Book Reviews: Emerging Technologies and Digital Communication
Topics: Document design, structured authoring, content strategy, and data/visual communication
Privacy and Security for Large Language Models: Hands-On Privacy-Preserving Techniques for Personalized AI
Baihan Lin. Sebastopol, CA: O’Reilly Media. 2026. 318 pages, including index.
Index terms — Artificial intelligence, cybersecurity, data privacy, federated learning, large language models, machine learning
Reviewed by Yogiraj Awati, IEEE Senior Member (yogirajawati@gmail.com)
Review posted July 1, 2026
In Privacy and Security for Large Language Models: Hands-On Privacy-Preserving Techniques for Personalized AI, Baihan Lin addresses a challenge that has grown increasingly urgent as organizations move beyond experimenting with large language models (LLMs) and begin deploying them in consequential settings: how to harness the capability of these systems while preserving data privacy and maintaining security. This book is written for AI practitioners, machine learning engineers, data scientists, and security professionals who are moving from controlled experimentation to production deployment and who encounter privacy and compliance requirements that standard machine learning texts rarely address in depth. Lin’s aim is to bridge the gap between theoretical privacy research and the practical implementation demands faced by practitioners. He succeeds in providing a structured, technically substantive guide that is grounded in code examples, real-world case studies, and domain-specific applications.
The opening chapters establish foundational concepts in LLM architecture and then shift to a systematic treatment of privacy and security risks. Lin introduces key metrics (differential privacy, attack success rate, false positive rate for membership inference, and reconstruction error for model inversion) and shows how these can be applied using his custom auditing framework, LLMPrivacySecurityEvaluator. This section provides a structured vocabulary and measurement approach that is often missing from general machine learning resources. The custom auditing framework allows teams to assess exposure before deployment rather than after an incident.
The core technical content spans privacy-preserving training techniques and secure deployment strategies. Lin covers differential privacy with stochastic gradient descent, federated learning architectures, homomorphic encryption, and multi-party computation, applying each specifically to the LLM context rather than treating them as general principles adapted loosely from database or classical machine learning literature. The deployment chapters address secure API design, authentication and authorization, network isolation, and model versioning and update protocols. A chapter on adversarial attacks (jailbreaking, embedding space attacks, prompt injection) describes both the threat landscape and mitigation approaches including red-teaming methodologies and robust fine-tuning. This consistent emphasis on implementation rather than theory will make this material directly applicable to current engineering problems for IEEE readers working in systems engineering, communications security, or applied machine learning.
The later chapters address ethical considerations and the broader sociotechnical environment in which LLMs operate. Lin examines bias and fairness in fine-tuned models, explainability under privacy constraints, and the tensions that emerge when privacy-preserving mechanisms interact with fairness requirements. A chapter on cultural, social, and legal landscapes covers copyright, data protection regulation, algorithmic discrimination, and liability. The book closes with applied case studies in healthcare and legal AI, demonstrating how the privacy-preserving techniques covered in earlier chapters translate into domain-specific architectures. These examples help ground abstract mechanisms such as differentially private clinical note analysis and federated learning across law firms in recognizable operational contexts.
A notable strength of the book is its treatment of trade-offs. Lin does not present privacy-preserving techniques as straightforwardly beneficial additions to existing pipelines. Instead, he consistently addresses the costs: differential privacy introduces noise that affects model utility; federated learning requires careful coordination of aggregation and introduces communication overhead; homomorphic encryption imposes computational demands that constrain deployment contexts. This framing will resonate with systems engineers who are accustomed to evaluating design decisions under constraint rather than selecting from an idealized menu of options. The hands-on code examples in Python throughout the book reinforce this practical orientation and allow readers to evaluate the presented techniques against their own deployment environments.
Readers who work primarily in resource-constrained or embedded environments may find that the deployment architectures assume access to cloud infrastructure and computational resources that are not universally available. The book’s case studies centered on healthcare and legal AI in enterprise settings reflect deployment contexts with significant organizational capacity. Practitioners in smaller organizations or in domains with different infrastructure constraints may need to adapt the presented frameworks. This is a minor limitation given the book’s scope and does not diminish the value of the technical content, but it does signal an opportunity for follow-on work focused on privacy-preserving techniques in lower-resource environments.
Privacy and Security for Large Language Models offer at least two compelling reasons for engineers and technical professionals to engage with it. First, it provides implementation-level guidance that addresses the specific characteristics of LLMs rather than repurposing techniques designed for simpler systems, filling a gap that practitioners have had to navigate through scattered research papers. Second, it connects privacy and security considerations to the broader ethical and regulatory environment, helping technical readers understand the compliance and governance dimensions of decisions that might otherwise appear purely engineering in nature. The combination of rigorous technical content and applied case studies makes the book a practical reference for anyone responsible for deploying LLM-based systems where data sensitivity, regulatory requirements, or adversarial exposure are serious concerns.
Human–AI Interaction and Collaboration
Dan Wu and Shaobo Liang, eds. Cambridge, United Kingdom: Cambridge University Press, 2026. 294 pages.
Index terms — Artificial intelligence, human–AI collaboration, user experience
Reviewed by Guiseppe Getto, IEEE PCS Member, Mercer University (getto_ga@mercer.edu)
Review posted July 1, 2026
Human–AI Interaction and Collaboration, edited by Dan Wu and Shaobo Liang, is a timely collection for readers trying to understand what happens when artificial intelligence (AI) systems shift from tools users operate to collaborators users must evaluate, trust, question, and sometimes resist. The book provides a broad, research-based account of human–AI interaction across domains while keeping user experience, privacy, credibility, and responsible design in view. For IEEE Transactions on Professional Communication readers, its value lies in mapping issues technical communicators increasingly face as AI enters knowledge work, documentation, search, content production, healthcare communication, and scientific discovery. The book is most useful to researchers, graduate students, and practitioners in human-centered interaction (HCI), information science, AI studies, health informatics, and communication. It is not a practitioner handbook, but it offers frameworks and examples that can inform AI design and evaluation.
The book’s main contribution is that it treats human–AI interaction as both an interaction and a collaboration problem. This distinction matters. Users must not only operate AI systems; they must decide when to trust them, how to interpret their outputs, and how much agency to grant them. In other words, AI does not simply add another interface. It adds another actor to the communication situation.
The book is organized into 12 chapters that move from broad frameworks to specific domains of practice. Chapter 1 introduces the need for human-centered AI, and Chapter 2 offers a framework for human–AI interaction and collaboration. This framework is especially useful because it treats trust, explainability, role assignment, and interaction boundaries as communication issues, not just technical ones. If users cannot understand what an AI system is doing, what role it is playing, or when to override it, the system may function technically while still failing in practice.
Chapters 3 and 4 address privacy and credibility in human–generative AI interaction. Together, they show that AI-generated information must be assessed, contextualized, and communicated responsibly, not simply produced efficiently. Chapters 5–7 focus on crowdsourcing, search, and misinformation, connecting AI to long-standing technical communication concerns about content quality, moderation, knowledge sharing, and information literacy. Chapters 8–10 move into collaborative intelligence, health information, and scientific discovery, showing that human–AI collaboration must be evaluated within specific tasks, domains, users, and organizational constraints. Chapter 11 addresses generative AI challenges, and Chapter 12 synthesizes the book’s key findings. The overall argument is clear: AI must become more humane, understandable, and accountable if it is to support human work.
For technical communication professionals, the book’s value is that it treats AI as part of a larger communication situation involving users, systems, tasks, information, and contexts of use. Several chapters speak directly to our field, especially those on privacy, credibility, search, crowdsourcing, and misinformation. The book also clarifies why human-centered AI matters: if users do not understand an AI system’s role, limits, outputs, or risks, then the system may work technically while still failing communicatively.
Scope is one of the book’s major strengths. The editors cover a wide range of issues without losing the central thread of human–AI interaction. Its grounding in information science and HCI also helps move the conversation beyond broad ethical claims toward more specific concerns: privacy identification, credibility assessment, user understanding, misinformation, and domain-specific collaboration. These are the kinds of issues technical communicators need to understand as AI becomes embedded in documentation workflows, research practices, and organizational decision-making.
Scope is also one of the book’s main limitations. Because the collection covers so much ground, some chapters read more like research summaries than fully developed models for practice. Readers looking for a unified methodology for designing human–AI collaboration will need to assemble one across chapters. Professional communication is also not the book’s main field of reference, so readers will need to translate its insights into documentation, user experience (UX), content governance, and cross-functional workflows. That translation work is manageable, but it is still work.
Ultimately, Human–AI Interaction and Collaboration is valuable because it takes seriously the idea that AI systems must be designed around humans, not merely deployed around them. It will be especially useful for scholars and graduate students working across technical communication, UX, HCI, information science, and AI ethics. For practitioners, it provides a research foundation for asking better questions about AI systems: What do users understand? What do they trust? What work is being shifted to the system, and what work must remain human?
Overall, Human–AI Interaction and Collaboration is a strong research resource for readers who want a broad, multidisciplinary view of human-centered AI. Its greatest value is not in offering a single design method, but in bringing together current research on the interaction problems AI creates: privacy, credibility, trust, search, misinformation, and domain-specific collaboration. The collection is uneven in places, as edited collections often are, but it gives technical communication scholars and graduate students a useful foundation for thinking more carefully about AI as part of knowledge work.
Untangling AI: Driving Business Success Through Enterprise Automation and AI Agents
Matt Kesby. Hoboken, NJ: John Wiley & Sons, Inc. 2026. 406 pages, including index.
Index Terms — Artificial intelligence, automation, business strategy, digital transformation, professional communication
Reviewed by Naveen Kumar Puppala, IEEE Senior Member (naveenp2708@gmail.com)
Review posted July 1, 2026
Most books about artificial intelligence (AI) get lost in the technology. They walk you through models, benchmarks, and the latest releases, and by the end you know more about the tools than you did, but you still have no idea how to apply any of them in practical work the next day. Untangling AI: Driving Business Success Through Enterprise Automation and AI Agents takes a different route. Matt Kesby wrote it for the people who actually have to make AI work inside an organization: business leaders, managers, consultants, and technology professionals. The result is a book that stays grounded in business value instead of hype, and that is exactly what makes it worth reading.
The central argument is simple: AI success depends on far more than the technology itself. Kesby builds his case around the AI Sweet Spot Framework, which covers four areas: Strategy, People, Technology, and Execution. I liked this structure because it matches what actually happens in real companies. Most AI projects do not fail because the technology is missing. They fail because nobody defined the problem clearly, the employees were not ready, or the execution fell apart somewhere between the pilot and production. Framing AI as organizational change rather than a software install is the right instinct, and the book holds onto it from start to finish.
The Strategy section is one of the strongest parts. Kesby pushes leaders to figure out where their organization actually stands before they start buying or building anything. That sounds obvious, but plenty of teams jump straight to which AI tool should we use? before they have asked what business problem they are solving. The AI Agility Check and the roadmap material give leaders a way to slow down and answer that question first. Kesby uses the Pirate and Navy comparison, which is my favorite idea, in this book. The Pirate side is about speed and experimentation; the Navy side is about structure, scale, and governance; and the point is that a healthy organization needs both at once. It is a simple image, but it sticks.
The People section was the one that stood out the most to me. Kesby does not pretend that AI adoption is painless. He talks directly about fear and resistance, which many business books quietly avoid. When the message employees hear is automation is coming for your job, anxiety is a reasonable response. The book works hard to shift that conversation toward amplification instead of replacement by putting real weight on psychological safety, upskilling, and bringing people into the change rather than having it done to them. This matters more than most leaders admit. A technically sound rollout can still collapse if the people affected feel excluded or threatened, and Kesby is honest about that risk.
The Technology section manages a tricky balance. It explains large language models, prompt engineering, context engineering, retrieval-augmented generation, application program interfaces, vector databases, and prompt injection in language a non-technical reader can follow, without oversimplifying it to the point where it loses practical value. What I appreciated most is that the book refuses to treat AI as magic. Kesby is clear that these systems still depend on data quality, sound processes, security, governance, and human review. His discussion of prompt injection and reliability is especially worth reading because it makes the case that organizations cannot simply trust AI outputs without controls in place.
Execution is where many AI ambitions fail to become real business outcomes, which the book treats it seriously. Kesby is direct that ideas alone produce nothing. Organizations need a repeatable path from spotting an opportunity to building, testing, deploying, training people, running it, and evaluating the result. The DDDEEE (Discover, Design, Deploy, Educate, Execute, Evaluate) Framework gives business teams a way to picture that path without needing an engineering background. I also found the Value Impact concept genuinely useful. It asks leaders to estimate what an AI opportunity is actually worth before committing to it, which is a quiet but effective way to stop the scattered, this sounds exciting experiments that drain budgets and fail to produce meaningful results.
A few things make the book easy to recommend. The language is plain and the examples are clear, so a leader who has never written a line of code can still follow the argument and make decisions with it. Another strength is the attention to communication, culture, and trust. AI adoption often succeeds or fails based on how well leaders explain the change, document the new processes, train their teams, and set expectations. Poor communication can make a perfectly good initiative feel confusing or threatening, and the book understands that. I also liked the practical take on AI agents. Rather than encouraging readers to build elaborate agents for their own sake, Kesby ties them back to real processes and outcomes: identify the problem, understand the workflow, pick the right level of automation, and measure what happens.
The book is not perfect; a few areas could be stronger. Some of the statistics and forward-looking claims are striking, but they would land harder with more context. When the book cites adoption rates, productivity gains, valuation effects, or failure rates, I wanted to know the source and the scope. Are these numbers drawn from large enterprises, from small and medium businesses, or from a mix of everything? Without that, some of the strongest claims would be more persuasive with additional support.
I would also have liked more worked examples. Several of the frameworks are genuinely useful, but a reader trying to apply them for the first time would benefit from seeing one filled in completely. The AI Agility Check, the maturity scale, the Three-Horizon Framework, and the Value Impact formula would all be stronger with a single example carried from start to finish: the problem, the opportunity, the calculation, the risks, and the expected business value. One clean walkthrough would make these tools far easier to put into practice.
The last thing I would add is more on governance and risk once AI moves into sensitive workflows. The book does cover security, reliability, and responsible adoption, but a clearer checklist for privacy, bias, human approval, auditability, and accountability would help. That gap matters most in regulated industries such as healthcare, finance, insurance, and government, where a wrong AI decision carries real consequences.
None of that takes away from the overall value. Untangling AI is a timely, practical guide for anyone trying to think clearly about enterprise AI adoption. Its real contribution is connecting strategy, people, technology, and execution in one coherent framework instead of treating them as separate problems. It is especially good for practitioners who are tired of the hype and want a grounded place to start. The book keeps returning to one idea: successful AI adoption is not about picking the newest tool. It is about choosing the right problem, preparing your people, building trust, managing risk, and measuring results.
I would recommend it to business leaders, technology managers, consultants, and anyone working on digital transformation or automation. Practitioners will get usable frameworks for planning their AI work. Readers who care about communication, culture, and organizational change will find plenty here too, because the book shows how much those factors shape whether technology actually gets adopted. Its strongest message is also its simplest: AI transformation is a human and organizational challenge at least as much as a technical one. That is what makes the book relevant, and that is what makes it worth reading.
Leading Enterprise AI Programs: Optimize AI Teams for Value Creation
Patrick Bangert. London, UK: Kogan Page. 2026. 224 pages, including index.
Index Terms — Artificial intelligence, data governance, enterprise AI, technology leadership
Reviewed by Lokesh Karanam, RR Donnelley (lokeshkaranam3@gmail.com)
Review posted July 1, 2026
Patrick Bangert’s Leading Enterprise AI Programs: Optimize AI Teams for Value Creation is mainly for executives and senior practitioners who build and run artificial intelligence (AI) capability in an organization, including chief AI and data officers, program and portfolio leads, and engineering and product managers who must turn AI ambitions into durable business outcomes. Data scientists, machine learning engineers, and governance or risk professionals form a natural secondary audience, as do consultants advising on AI transformation. The book sets out to describe how to establish and operate an agile, ethical, and business-focused AI program for the whole enterprise rather than a scattering of disconnected pilots. Overall, it largely achieves this purpose: it offers a coherent operating model and a practical vocabulary for treating AI as a managed portfolio of value-generating use cases rather than a series of one-off experiments.
Bangert’s central argument is that the difficult problems of enterprise AI are organizational and managerial, not merely algorithmic. He contends that lasting value comes from choosing and prioritizing the right use cases, standing up an operating model suited to the organization’s goals, and cultivating a community of practitioners. A recurring theme is that AI work should be anchored to clearly framed business problems and to the users and interfaces through which the work is actually consumed. The author supports these claims with the structure of a practitioner who has run such programs, weaving strategy, governance, and technical foundations into a single arc. The framing that organizations should bring AI to the data rather than moving data to the AI captures the book’s bias toward pragmatism and toward fitting AI into existing enterprise realities.
Methodologically, the book is grounded in lived executive experience rather than formal empirical study. Bangert draws on more than two decades of leading AI and machine learning teams across industry, and the guidance reads as distilled practice heuristics, operating patterns, and cautionary observations rather than as hypotheses tested against controlled data. This situates the work firmly within practice-oriented professional guidance, which suits its intended readers, who need defensible decisions more than they need citations. The book’s concerns are, at their core, problems of professional communication: framing an AI use case so stakeholders understand the value at stake; persuading executives to fund a portfolio rather than a demo; negotiating governance and ethical guardrails across legal, security, and business functions; and communicating model performance and risk to audiences who do not speak the language of statistics. Bangert’s emphasis on operating models and community-building is, in effect, an argument about how an organization socializes and sustains its technical work.
The book’s treatment translates abstract leadership advice into decisions a reader can take on Monday morning, focusing on use-case selection and prioritization, the operating model and its roles, and the governance scaffolding needed to deploy AI responsibly. The attention to citizen data scientists and to user-facing interfaces is a welcome corrective to AI writing that stops at the model boundary; it keeps adoption and consumption in view. For this reason, the book lends itself well to applied use: chapters could anchor executive education modules, AI program-management workshops, or onboarding for newly formed centers of excellence, where its frameworks give participants a shared structure for messy, cross-functional work.
The book’s limitations are largely matters of scope and timing. It generalizes across industries, but some of its guidance necessarily stays at the level of the operating model and may need translation before it fits a specific regulatory or technical context. Readers in highly constrained domains will have to supply the particulars. It leans on experience, not hard evidence, so readers will need to treat it as practical insight rather than theory. The field also moves fast, which is a built-in risk. A 2026 book about AI programs must keep its principles useful even as the technology continues to shift. The strategy and governance advice ages slowly; the tooling specifics will not, so weigh your reading accordingly. These are not so many flaws as scope boundaries; the book is most powerful when read as a seasoned leader’s playbook for a defined managerial problem.
On balance, Leading Enterprise AI Programs is a clear, well-organized, and credibly experienced account of how to make enterprise AI deliver value rather than slideware. Explicit mapping of its operating model and governance frameworks to adjacent disciplines, such as data engineering, technical communication, and risk management, would help readers outside the core AI leadership audience adapt the guidance to their own roles. AI and data officers, program and portfolio managers, and their technical leads will get the most from it. Educators and consultants will find it a handy supplementary text for showing how AI really gets governed, funded, and communicated in large organizations.
The AI Matrix: Profits, Power, Politics
Daniel Mügge, Regine Paul, and Vali Stan. Newcastle upon Tyne, UK: Agenda Publishing. 2026. 160 pages, including index.
Index Terms: Artificial intelligence, global political economy, labor, technology policy
Reviewed by Sriharsha Makineni, IEEE Senior Member (sriharsha.makineni@gmail.com)
Review posted June 24, 2026
The AI Matrix: Profits, Power, Politics is a clearly written book about artificial intelligence (AI) as an economic and political force. Unlike many other AI books that treat this subject as primarily a technical breakthrough or portray its impact solely by progress versus danger, Daniel Mügge, Regine Paul, and Vali Stan frame their inquiry differently by asking far more grounded questions: Who stands to benefit from AI? What are the costs associated with AI? How are these outcomes shaped by companies, governments, workers, and global power structures? Readers interested in technology policy, professional communication, engineering ethics, political economy, and AI governance will find this book useful. It provides a valuable framework for practitioners working within the context of AI systems who wish to have a greater understanding of the potential social and economic implications surrounding AI.
A primary strength is the unwillingness of the authors to take the typical AI debate at face value. Much of the general public’s conversation surrounding AI focuses on the extremes of “AI will save or ruin us.” However, the authors move beyond this simplistic framing by suggesting that AI is influenced by profit-driven decision-making, political choices, public perception of AI, and international competitive pressures. While this may seem like a straightforward concept, it is one that is frequently absent from the majority of mainstream AI writings.
In Chapter One, the authors discuss how one should use a political economy lens to study AI. They show that large corporations, investments, public policies, and national competitions have a heavy influence on AI’s development. Additionally, they demonstrate how AI stories can affect funding, regulations, hiring, purchasing and public trust. This aspect matters in professional communication because organizational description of AI shapes how stakeholders (users, policymakers, employees, investors) perceive technology.
Chapter Two expands upon AI by showing that there is no singular form for it. Rather, AI encompasses the types of tools, systems, infrastructures, and labor practices. This is an important distinction because many public conversations about AI focus on Generative AI while ignoring other types of AI (surveillance systems, robots, logistics tools, predictive analytics). The authors highlight that AI relies on hardware (chips), software (data centers/cloud computing), content, and often hidden human labor. Thus, taking a broader view of AI enhances the book over many commentaries regarding AI that focus on chatbots and model performance.
Mügge et al. proceed to analyze how geopolitical competition influences AI and examines how it affects various industries differently. Their examination of U.S. efforts to develop AI capabilities and the challenges presented by China is one of the book’s most effective sections. Rather than simplifying the dynamic into a binary (“U.S. vs. China”), the authors show how states, firms, infrastructure, and regulations interact. They also provide substantial details regarding European attempts to create alternative regulatory frameworks for AI, while acknowledging that Europe remains reliant on larger technological powers. Thus, the authors present a more nuanced assessment than would typically occur with a standard innovation narrative.
The chapters related to sectors and labor are robust much like the previous section. Rather than simply examining if job losses result from AI’s adoption, the authors explore a more productive line of inquiry: how will AI transform the nature of work (organization structure, quality of work, level of control afforded to workers, value derived from work). This represents an improvement over previous research inquiries, since job loss is certainly important, but so are issues like work intensification, deskilling, platform dependence, and erosion and strengthening of worker power. The authors are strongest when considering labor as a human and social phenomenon rather than merely as an input in a production function equation. This broadens the book’s relevance beyond economists and policymakers to include engineers, managers, educators, and professional communicators who develop, design, and explain AI-enabled workplace systems.
Of further value is the book’s attempt to include perspectives from outside of the U.S–-China–Europe axis of power. Many books concerning AI reference themselves as global yet continue to exclusively address only a handful of developed regions. The AI Matrix presents a more successful effort to examine disparate impacts experienced globally (dependency, infrastructure gaps, outsourcing data work, local capacities to develop AI in nations which rely on foreign-based platforms, and capital) regarding inequality. Globalization is among the book’s clearest contributions, showing how AI can reproduce historical economic inequalities in new technical forms.
Technical professionals and researchers can expect to see technical (model architecture) and non-technical (policy implications) material discussed related to AI. However, since the focus is on understanding what drives trends in AI and the role that societal factors play in shaping them, the book will be too broad to provide depth on either topic. Technical professionals, as well as researchers interested in the development of new AI applications, should be able to gain insight into the social context in which such applications exist. The rapid pace of change within the AI field also means that examples used may become quickly outdated . Therefore, the authors might consider including summary tables or other types of visual framework at the end of each chapter to help users identify key points and make connections across topics.
These limitations reflect several challenges associated with producing a book about a rapidly changing area. The success of the “AI Matrix” lies in its ability to give readers an approachable way to think critically about AI’s future (p. 2). Importantly, it clearly states that AI’s future is not set. Rather, it is dependent upon numerous public choices made by governments, business incentives, worker resistance to automation, regulatory actions taken by governments, and global inequality. This is an important message when many organizations view AI adoption as inevitable.
There are two primary reasons you will want to read The AI Matrix. First, it provides non-expert readers with a good understanding of the economic and political influences that drive AI. Second, it provides experts and technical readers and writers with a more suitable vocabulary to ask socially-responsible questions regarding AI’s deployment including: Who owns or controls the AI system? Who gains benefits from the AI system? Who gets monitored through the AI system? And who is excluded from using the AI system? For these two primary reasons, The AI Matrix represents a solid contribution to ongoing conversations related to AI, power, and society.
Governing Pandora: Leading in the Age of Generative AI and Exponential Technology
Andrea Bonime-Blanc. Washington, DC: Georgetown University Press. 2026. 288 pages; including index.
Index Terms — Artificial intelligence governance, corporate governance, emerging technologies, risk management, technology leadership
Reviewed by Lokesh Karanam, RR Donnelley (lokeshkaranam3@gmail.com)
Review posted June 24, 2026
Andrea Bonime-Blanc’s Governing Pandora: Leading in the Age of Generative AI and Exponential Technology is aimed at senior organizational leaders (executives, board directors, and the chief risk, ethics, compliance, and technology officers who advise them) and the policymakers who must guide institutions through swift technological change. Readers will find its integrative treatment of exponential technologies useful in corporate governance, enterprise risk management, and sustainability fields. One guiding question drives the book: what does it mean to lead responsibly in the age of exponential technology, a period Bonime-Blanc dubs the Age of Pandora. She offers a roadmap for governing generative AI and other fast-advancing technologies in a way that captures their promise and protects against their perils. The book provides a set of practical frameworks that leaders and organizations can apply to their own settings.
Bonime-Blanc advances several connected claims: that exponential technologies are converging and accelerating faster than existing governance can manage; that a clear situational awareness of the broader technological landscape is a prerequisite for responsible leadership; and that organizations do not need a single policy, but a durable governance mindset spanning culture, ethics, risk, and strategy. She uses a four-part architecture to develop these claims. Part I maps a global technological megascape and a set of technology-related megatrends, ranging from the challenge of governing the seemingly ungovernable to the geopolitical and socioecological dimensions of technology. Part II introduces the book’s centerpiece, a five-element exponential governance mindset built from leadership, ethics, impact, resilience, and foresight, while Part III operationalizes it through a set of organizational tips and trustmarks. The argument is supported less by formal data than by conceptual models, tables and checklists, illustrative cases, and a distinctive coined vocabulary, which together give the framework structure and a shared language.
Methodologically, the book is not based on empirical research but on Bonime-Blanc’s own decades of advisory practice in governance, ethics, and strategic risk. She writes as a practitioner-strategist on the basis of client-facing experience, current events, and a synthesis of governance thinking to build normative frameworks, not to test hypotheses. This places the book in the realm of practice-oriented leadership guidance, which is well-suited to readers who require actionable orientation in an uncertain environment. At its heart, it is about professional communication.
The book employs many coined terminologies, which are intentional, but can come across as jargon, which may require a level of understanding before its value becomes apparent. Because the frameworks are based on practice rather than tested by research, scholars may find this book more useful as a rich source of practitioner insight. The focus on large, board-governed, and mostly Western organizations means leaders in smaller, public-sector, or non-Western contexts may need to adapt the guidance to their own situations. There are fewer flaws than there are scope boundaries. The book is best read as a strategic field guide for senior leaders and not as a technical manual or academic monograph.
Governing Pandora serves as a candid, systematic, and thorough narrative of what responsible leadership demands in a period of accelerating technological change. A practical improvement would be to translate its five-element mindset explicitly into governance and risk standards that readers already use, such as the NIST AI Risk Management Framework or existing enterprise-risk and ethics models, so that organizations can adopt and not reinvent. As a secondary text, it would serve educators who want students to wrestle with real stories of technology governance, leadership, and stakeholder communication. Executives, board directors, technology officers, and policymakers will find this book useful for its practical framework, its tools for building situational awareness and a responsible tech culture, and its core argument that governing powerful technology is ultimately a leadership and communication challenge.
AI in Business For Dummies
Jeffrey Allan. Hoboken, NJ: John Wiley & Sons. 2026. 320 pages, including index.
Index Terms — AI adoption, data governance, generative AI, human-AI interaction, professional communication
Reviewed by Adarsh Mittal, Nvidia (adarshm9031@gmail.com)
Review posted May 27, 2026
Coming from a background in large scale system design and engineering environments, where communication gaps can directly translate into execution issues, I read AI in Business For Dummies with a slightly different lens. I was not only looking for whether it explains artificial intelligence (AI) concepts well, but also whether it helps bridge understanding across technical and business roles. The book clearly aims to lower the barrier to entry for AI in business contexts and, in many places, it succeeds in making the topic approachable.
One of the book’s strengths is how Jeffrey Allan breaks down AI into approachable concepts without assuming a strong technical background. He writes in a style that the general audience can follows without difficulty, such as the sections explaining the different AI types: supervised, unsupervised, and reinforcement learning. The examples, in areas like retail fraud detection or recommendation systems, help ground the discussion. That said, the simplification sometimes comes at the cost of precision. For instance, while Allan introduces the categories clearly, there is limited discussion on tradeoffs, failure modes, or system boundaries. From a communication standpoint, this matters because oversimplified explanations can lead to misaligned expectations between business leaders and technical teams.
This book is more effective when it discusses practical workflows for integrating AI into business processes. The step-by-step breakdown covers identifying needs, selecting tools, preparing data, and iterating through pilots is useful, and reflects a reasonable deployment lifecycle. The emphasis on starting small and scaling gradually is particularly relevant, since many organizations underestimate the complexity of system integration. However, even in these sections, the communication aspect is implicit rather than explicit. For example, while the book advises aligning AI initiatives with business goals, it does not go deep into how these goals should be documented, communicated across teams, or translated into measurable requirements that engineering teams can act on.
Another area where the book adds value is by illustrating real-world applications across functions like marketing, customer service, and finance. The examples around personalized email generation, automated support systems, and financial reporting show how AI can reduce manual effort and improve efficiency. These examples are relatable and help non-technical readers understand potential use cases. At the same time, they are presented mostly as success stories. There is less discussion of failure scenarios, integration challenges, or situations where AI outputs degrade decision quality. In practice, communicating both capabilities and limitations is essential, especially when deploying systems that influence customer interactions or financial decisions.
The sections on metrics and evaluation are useful but are written at a surface level. The book identifies key measures that are reasonable starting points, like time savings, accuracy, and customer satisfaction. However, Allan offers limited guidance on how to track, validate, or communicate these metrics to different stakeholders. For example, a metric that is meaningful to leadership may not be actionable for engineering teams. This gap again reflects a broader pattern in the book where operational communication layers are not developed.
One part that stands out more positively is the discussion around tool selection and evaluation. The comparison between free and paid tools, and the emphasis on reliability, security, and integration, is practical. The suggestion to treat early adoption as experimentation rather than commitment is realistic and aligns with how most organizations evaluate new systems. Still, the decision-making process is framed mostly at an individual or small-team level. In larger environments, tool selection requires structured communication, review cycles, and alignment across multiple stakeholders, which Allan does not fully address.
The coverage of data privacy, security, and compliance is another important inclusion. The book highlights risks associated with handling sensitive data and stresses practices like encryption, access control, and regulatory awareness. This is a necessary topic, especially given the regulatory landscape around AI. However, similar to other sections, the treatment is more checklist-driven than communication-driven. For example, while it advises documenting processes and maintaining compliance records, it does not explore how these should be communicated across legal, engineering, and business teams to ensure consistent understanding and enforcement.
A recurring theme throughout the book is that AI is positioned as an enabler of efficiency and innovation. The discussion on transforming business operations, improving decision making, and increasing productivity reflects current industry narratives. These claims are directionally correct, but the framing tends to lean optimistic. There is limited attention to organizational friction, change management challenges, or communication breakdowns that often occur during AI adoption. In real systems, these factors can have as much impact as the technology itself.
Overall, AI in Business For Dummies serves well as an introductory guide for business audiences who want to understand what AI can do and where it can be applied. It is accessible, structured, and covers a broad range of topics without overwhelming the reader. At the same time, from a professional communication perspective, it leaves some gaps. The content would benefit from more concrete discussion on how to communicate AI concepts across roles, how to structure decision making processes, and how to document and align system behavior with business expectations. As it stands, it is a good starting point, but it does not fully bridge the gap between conceptual understanding and effective communication in real organizational settings.
Reader Bot: What Happens When AI Reads and Why It Matters
Naomi S. Baron. Redwood City, CA: Stanford University Press. 2026. 268 pages, including index.
Index Terms — AI-as-reader, education, reading
Reviewed by Charlotte A. Weddington, a technical communicator at Hunter Douglas in Broomfield, CO 80020 USA (technicallyiwrite@gmail.com)
Review posted 15 April 2026
Naomi Baron’s book, Reader Bot: What happens when AI reads and why it matters, is a well-researched, informative, and engaging book that focuses on the contrast between what and why humans read versus what goes on when humans allow AI software to “read” for them. It is written from an educator’s perspective and is relevant to a broad audience—namely, anyone who reads and uses (or contemplates using) AI publicly-available software such as OpenAI’s ChatGPT, Anthropic’s Claude, and Microsoft’s Copilot. The author does a wonderful job of explaining the basics of generative AI, while asking thought-provoking questions for the reader to consider. Although the book is not about professional communication, Baron does discuss the consequences of using AI as a communication tool.
The author breaks the book into two parts containing seven chapters: What Counts as Reading; Who Counts as a Reader; What’s in the Library; Passing Judgment; Obligations: Schools and Salaries; Choosing to Read; and Reading as Social Connection.
In the first two chapters, Baron poses some intriguing questions about what reading is. For humans, does reading a tweet or a chat count? Do audiobooks count? A 2021 UK YouGov adult poll found that 53% did not think that listening to an audiobook was reading the book; 29% thought that it was; and 18% were undecided. What about reading a printed copy versus reading on a digital screen? When she surveyed 1,000 secondary and university students, she was surprised at the number who said reading the printed copy felt more authentic. No students described reading on a digital screen as real.
She also describes the difference between how humans read and how AI software reads. We read word by word, and scan ahead as we incorporate what we are reading into sentences, paragraphs, and chapters. We try to make sense of what we are reading by using experiential, contextual, and emotional memories. On the other hand, AI software relies on statistical analysis to make connections. It converts text into numerical representations (“tokens”) and then uses advanced predictive analysis to determine what word/token is most likely to come next (pg. 8). It simulates understanding without having the comprehension that humans do.
Who counts as a reader? Does it matter how much you read? How often? What kinds of material? For what purpose? To answer these and other questions, Baron discusses distinct types of readers, whether reading is obligatory or voluntary, and whether society reads well or not.
The third chapter contains some interesting and disconcerting information on how Large Language Models (LLMs) are fed and what is good data. For example, someone can ask ChatGPT to summarize a Supreme Court decision, and it creates an essay. Then the essay gets fed back into the LLM for future use. When the next person enters a similar request for the same Supreme Court decision, the response may be a bit watered down from the initial essay. After several requests for the same information, the output might not even look like a Supreme Court decision any longer. This theorized outcome is called “model collapse.” For example, the researchers who proposed this theory started with text about medieval architecture. By the ninth repeated generation, the word architecture was still there, but the rest of the output talked about “jackrabbits with differently colored tails” (p. 78).
Chapters four and five explore AI hallucinations and how AI software is now being used in various environments to scan resumes and college applications, read screenplays, and analyze academic manuscripts.
AI hallucinations are AI-invented facts, explanations, or references. This occurs when there is missing or ambiguous information in the LLM, which causes the AI to guess or statistically predict its response, even when it has no basis in reality. As Baron puts it, “When AI purports to read (or, for that matter, write or draw or do math), it throws up simulacra that, with growing frequency, appear reasonable. If humans have not done the reading (plus sometimes research) themselves, they can’t evaluate the reasonableness…” (p. 95).
For colleges and universities using AI-as-reader in 2023, 71% used it for reviewing transcripts, 73% for reading letters of recommendation, and 60% for reading essays. On one hand, high school students are concerned that their essays might not be read and analyzed by a person. On the other hand, students may be getting more help from AI in authoring their essays today; some reviewers no longer assume that essays are an accurate gauge of a student’s own writing ability.
In the last two chapters, Baron looks at the reasons humans read voluntarily, and at reading as a social activity. Some of the reasons that humans read voluntarily include reading for pleasure, to learn something, and for personal growth. One study found that what adults like most about reading includes entertainment, relaxation, and a way to escape from reality. Another study in 2024 by the National Literacy Trust found that over 50% of 8-to-18-year-olds agreed that reading helps them learn.
Regarding social connection, Baron writes, “Now that AI can write and read, we need to weigh possible outcomes. Will AI support the ways we connect with each other through the written word, or will it depersonalize this form of human communication?” (p. 164). For example, a research team found that when someone receives a response that is AI-generated, they assign negative traits to the sender such as being “less cooperative and more dominant” (pg 169). They cautioned that, “the influence of AI on human emotional communication is deeply concerning given that AI is writing billions of emails for us every day” and that we need to be aware that “an AI system designed to help people can have unintended social consequences” (p. 169).
“Reeling in Reader Bot,” the last section of the book, cites several studies regarding critical thinking. One research study showed that younger students who depended more on AI showed lower critical thinking ability than other groups. Another large-scale study of more than 4,000 participants, conducted at the University of Pennsylvania’s Wharton School, showed that using AI reduced students’ understanding of topics, which resulted in less perceptive essays.
This section also discusses and asks questions about the future of evaluations (letters of recommendation for employment, college applications, and peer reviews), interpersonal written communication (emails and letters), and AI technology. Regarding technology, Baron notes that AI will continue to evolve, and that we can expect a reduction in AI hallucinations. She also writes that AI companies are developing AI reasoning models, which excel in higher level reasoning and produce a long internal chain of “thoughts” before answering a prompt. However, she makes the caveat that it is not known whether some novel type of conceptual model will replace today’s LLMs, or how that innovative technology may affect how AI reads.
I found Reader Bot to be insightful about reading and how it contributes to overall human development. Baron raised many relevant questions about the consequences of AI-as-reader. I also found her views about generative AI software to be reasoned and balanced. I highly recommend this book for anyone who wants to thoughtfully consider when and how much to rely on AI for reading, analyzing, and writing.
AI for Knowledge
Andrew Cox and Mike Thelwall. Abingdon, Oxon: CRC Press. 2025. 84 pages, including index.
Index Terms – Artificial intelligence, governance, knowledge, libraries, research
Reviewed by Rahel Anne Bailie, Content Seriously (rahel.bailie@content-seriously.com).
AI for Knowledge is a book intended for students, researchers, librarians, and professionals who want to expand their perspectives on the application of artificial intelligence (AI) and generative AI (GenAI) in a range of domains. This book explores AI as a socio-technical system within a knowledge society. The suggestion that access to the benefits of AI as a human right is reinforced throughout the book, which informs both the tone and the examples used.
The book is helpful to professionals whose AI exposure may be focused in one area. The range of examples, while limited in such a short book, do cover a wide range of applications that helps spark the imagination for AI uses . Simultaneously, Andrew Cox and Mike Thelwall integrate discussion of environmental, social, and governance implications of using AI to bring knowledge to users in multiple contexts.
AI for Knowledge explores multiple ways that AI is being used and discusses both the benefits and dangers of the impact on society. This is particularly stressed as the impact on society, as access to information becomes more routed through AI models, which can result in digital inequalities. Digital inequalities result from an reinforce other inequalities. Whether the benefits of AI can be sustained is a theme that is woven throughout the book, with mindful, ethical use of AI becoming the overarching thrust of the work.
Given the current AI hype cycle where the industry progression from the peak of inflated expectations to the trough of disillusionment is moving fast, this book has kept its arguments sufficiently elevated to retain its relevance. This review is being written a year after the book’s publication, though the examples of AI and GenAI used in the book now seem quite modest. The examples are not meant to be exhaustive; they focus on areas of interest to communications professionals in their intended audiences.
AI for Knowledge is not about knowledge management in the way a typical knowledge manager may anticipate. Rather, the book covers multiple aspects of how information is created, disseminated, and understood.
Some of the more recognizable repercussions may now seem obvious, such as how the environmental impact of hardware mining and e-waste are felt disproportionately in the majority world. Other examples address more nuanced repercussions, such as the use of transcripts to analyze oral histories rather than recordings because of the loss of meaning and context gained from “close listening” to the original sound of a person’s voice (p. 39).
The book provides many use cases, alternating between research findings and examples as a way of bringing the theoretical to life. For instance, the discussion of the use of AI to assist with research evaluation is framed as a potential benefit, while the example of authors tailoring abstracts for high ChatGPT scores is framed as a potential limitation.
An area of interest is about the potential impact of AI on libraries and archives. These institutions are situated as key social institutions that promote access to knowledge and freedom of expression for all, while acknowledging the constraints on libraries in different contexts—for example, by available resources or by country. The potential of AI to democratize access to information and to broaden the available information to include the perspectives of marginalized communities can improve the connection of the broader population to library collections.
Whether discussing how AI could be used in an academic setting, in general or academic research, in libraries, or by government, those discussions are tempered by pointing out associated risks and limitations. The limitations are not meant to discourage but to show the importance of taking a considered approach.
This book’s ideal audience is the busy professional who wants a concise, rapid ramp-up to a rounded look at AI. The book is meant to provide a glimpse into the potential of AI for anyone wanting some applied uses of the technology. Simultaneously, the book draws attention to the possible negative repercussions of applying AI without consideration.
Academics will be able to use the information to build a plan for ethical use of AI for teaching, research, creation, review, or publication purposes. Cox and Thelwall illustrate how AI-assisted reviewer selection can match article keywords to potential reviewers even when reviewers have no matching keywords in their profiles, and how AI could help with summarising reviews to support editorial decisions. There is a suggestion that GenAI might also rate the quality of academic research; existing studies have shown that scores correlate positively with human scores for the same articles. The motivation is two-fold: first, to support efficient processing of papers; and second, to conserve the time and effort of subject matter experts with a shortage of time and energy.
The information is not meant as a comprehensive study of AI, but as a jumping-off point for readers to use as a foundation for further investigation. Whether the intent is to use AI as a peripheral tool in a practitioner’s toolkit or to develop novel uses for AI, AI for Knowledge is well-positioned to spark ideas.
Governing the Machine: How to Navigate the Risks of AI and Unlock Its True Potential
Ray Eitel-Porter, Paul Dongha, and Miriam Vogel. London, UK: Bloomsbury Business. 2025. 366 pages, including index.
Index Terms — Artificial intelligence, governance, regulatory compliance, risk management, trustworthy AI
Reviewed by Rajkumar Kuppuswami, IEEE Senior Member (rajk2888@gmail.com)
Review posted 15 April 2026
We’ve all heard the argument that artificial intelligence is moving too fast to regulate. It’s a common refrain in technology circles, often used to justify a deploy first, fix later mentality. But in Governing the Machine: How to Navigate the Risks of AI and Unlock Its True Potential, authors Ray Eitel-Porter, Paul Dongha, and Miriam Vogel push back hard against that idea. They argue that AI doesn’t exist in a legal vacuum at all. It’s already bound by existing laws covering civil rights, consumer protection, and product liability. The book is aimed at C-suite executives, board members, and the technical leaders actually building these systems as well as policymakers, compliance officers, and anyone studying technology ethics. The authors set out to provide a practical blueprint for building an AI governance program from the ground up. They manage to take a topic that usually feels overwhelming and break it down into something manageable and actionable.
The book is split into four main sections with some genuinely useful appendices. The authors make the case that AI governance isn’t just a compliance exercise; it’s a core leadership responsibility. Andrew Ng’s foreword really sets the tone when he points out that AI is a general-purpose technology like an electric motor. An electric motor isn’t inherently safe or dangerous; it depends on whether you put it in a blender or a weapon. Ng argues that we need to govern the application, not the underlying math. It’s a smart framing device that the authors carry through the rest of the text.
One of the most useful things in the book is how it categorizes AI risk. Instead of just talking about risk in the abstract, the authors break it down into nine specific buckets: accuracy, fairness, explainability, accountability, privacy, security, intellectual property, workforce impact, and environmental sustainability. They don’t just list these out; they dig into the messy realities of each one. When they talk about fairness, for example, they get into the weeds of how hard it is to actually define fairness mathematically across different demographic groups. They also spend time on automation bias, which is a dangerous tendency we have to just trust whatever the computer tells us. On the security front, they look at modern threats like prompt injection and data poisoning, treating them as serious enterprise cybersecurity issues rather than just theoretical academic problems.
Once they’ve laid out the risks, the authors pivot to how you actually build a governance program to manage them. They lean heavily on established frameworks like the NIST AI Risk Management Framework and ISO 42001. I appreciated their focus on accountability, as they argue for a cross-functional approach, bringing legal, compliance, and engineering teams to the same table. They even walk through how to use a RACI matrix to figure out exactly who is responsible for what. The book also gets into the mechanics of human oversight, breaking down the differences between having a Human in the Loop, Human on the Loop, and Human in Control. They also provide solid, practical advice on how to run Algorithmic Impact Assessments before a system ever goes live.
The third section is where Governing the Machine really stands out from other texts: it takes a truly global view of AI regulation. The authors spend a good amount of time unpacking the EU AI Act, explaining its risk-tier system in a way that’s easy to follow. They also look at the fragmented, agency-driven approach in the US, and then zoom out to cover what’s happening in over twenty other countries, including China, Brazil, and several African nations. If you’re working in a multinational organization trying to figure out how to comply with a dozen different sets of rules, this section alone is worth the price of the book.
For IEEE readers, engineers and technical communicators are the ones actually designing, deploying, and documenting these systems. We’re the ones who have to figure out how to explain a model’s decisions to a user or an auditor. Governing the Machine provides the ethical and regulatory context we need to understand why that work matters, and what the legal stakes are.
If I have a critique, it’s that the book is clearly written with large enterprises in mind. A startup or a small mid-market company might look at the governance structures proposed here and feel completely overwhelmed. In future editions, I’d like to see the authors offer a more tiered approach, such as a “minimum viable governance” framework for smaller teams that scales up as the company grows. Also, if you have a deep technical background, you might find their discussion of specific tools like SHAP values or differential privacy a bit surface-level. But given that they’re writing for a broad audience, that’s an understandable tradeoff.
Overall, Governing the Machine is an excellent book. It bridges the gap between high-level ethical theory and on-the-ground engineering practice. The appendices cover standards, legal frameworks, and training requirements that make this book a great reference guide to keep on your desk. Two reasons why any practitioner should read it: first, the nine-part risk taxonomy is something you can take and use in your next project meeting right away; and second, the global regulations breakdown is the clearest summary of the current legal landscape I’ve seen. If you want to move past treating AI governance as a roadblock and start using it to build better, more trustworthy systems, this is the book to read.
A History of Fake Things on the Internet
Walter J. Scheirer. Stanford, CA: Stanford University Press. 2025. 242 pages, including index.
Index Terms — Digital fabrication, digital fakery, disinformation
Reviewed by Diane Martinez
Review posted 15 April 2026
One thing the public is keenly aware of in today’s manufactured digital world is the possibility that what we see, hear, smell, taste, or touch is fake. Deception has always been part of technological advances, but the internet spawned new levels of fakery and deception that has many people on edge and skeptical about most anything they encounter digitally or in real life. At times it can feel like we are living in a completely fabricated world or that our identity or some other aspect of our lives will be stolen right before our eyes. But are things really as bad as they seem? Walter J. Scheirer in his book, A History of Fake Things on the Internet, claims they are not.
Scheirer’s approach to talking about a subject that scares and frustrates most people: being deceived is one unique aspect in this book. It is not an exposé on fakery or a dangling carrot of “come see what you didn’t know was fake.”Instead, Scheirer admits that most of what we share has been altered in some way and the “bulk of users on various social-media platforms are bots” (p. 20), but still, “things aren’t nearly as bad as they might seem” (p. xviii).
Scheirer says he is not interested in being an alarmist, so his book is about “technological trends and individual incidents that have shaped the development of labyrinthine alternate realities on the Internet” (p. 23). He explores in eight chapters these trends and incidents that offer abbreviated histories of four influential communities: computer hackers, digital artists, media-forensics specialists, and AI researchers.
The introductory chapter. “Restyling Reality,” explains the need to look at historical evidence of fakery beginning with humans’ obsession of transcending the physical world and the significance of myth in society.
The “On the Virtual Frontier of the Imagination” chapter tunnels into the underground world of hackers to help dispel the idea that anything fake was created with malicious intent and what it means to search for the truth.
Part of Scheirer’s story entails looking into the power of digital manipulation in “Photoshop Fantasies.” This chapter is not a rehash of PhotoShop’s manual; instead, it is a study into how image manipulations affect societies through historical and “alternate” realities. He discusses insightfully case studies on how governments around the world have altered historical images to tell history through a particular lens and how political and social memes provide their own historical narratives.
“Cheat Codes for Life” is one of the book’s most technical chapters because it dives into the depths of the hacking scene, cheat codes, the genesis of blackhat and whitehat hackers, and computer security.
“[F]orensic analysis of images can be helpful for rooting out harmful information,” (p. 109) and is the subject of “Speculative Sleuths.” The idea that manipulated images could become so real that it would be impossible to tell digital fakes from real images is a major societal concern. Digital fakes have historical, political, social, and even more disturbing, criminal, implications. Thus, this chapter introduces the Internet’s underbelly of child porn, war crimes, and fake science. Along those same lines, “Virtualized Horror” examines the allure of the horror genre and shock content on the Internet.
Along with humans’ interest in transcending the physical world is a fascination, and sometimes obsession, with knowing the future. “Dreams of Clairvoyant AI” scrutinizes the capability of AI to predict the future and the actors who are faking their way into fame by parading a scientific work as scientific. Scheirer also discusses the capabilities and limitations of other algorithmic forecasting like elections and climate change forecasting.
The “Creative Spaces” chapter ends the book ends on a positive note. Although each chapter is self-contained and does a thorough job of explaining the significance of the four communities Scheirer examined and the implications of the work within those communities, this last chapter is an excellent culmination of all the previous information. He follows through nicely on his comment that “a good history shouldn’t be compiled just to satisfy intellectual curiosity. It should contribute to our understanding of the present and help us make better decisions and avoid repeating the mistakes of the past” (p. 171). Thus, he addresses our relationship with technology and how to live well with what is yet to come. Despite the more disturbing information from previous chapters about bad things on the Internet, Scheirer still concludes that things are not as bad as they seem, but you may not really know whether that’s true or not until you read his book.
As far as readability is concerned, Sheirer avoids unnecessary technical jargon or explains it when it is necessary to use. Although the content of programming and the world of hackers was intimidating at first, overall, the book is easy-to-read and understand for non-technical audiences, but there are moments when those who are not familiar with computer science may have to slow down or reread some sections.
If you think this type of history is not really your thing, I highly suggest rethinking that stance because A History of Fake Things on the Internet is an excellent and thought-provoking history about issues that affect us all, even more than we might know. And although Scheirer concludes that things are not as bad as they seem, we should all be aware of how we got where we are now, so we are informed about decisions we make for the future.
AI Meets Strategy: A Product Manager’s Guide to Leading Innovation
Anshuman Srivastava, Abhinav Garg, and Anshuman Mishra. New York, NY: Apress Media LLC (Springer Nature). 2026. 670 pages, including index.
Index Terms — AI product management, digital transformation, professional communication, organizational strategy, responsible AI
Reviewed by Bhavya Lahari Bommireddy, Senior Software Engineer, Microsoft, (bhavyalahari1497@gmail.com)
Review posted 25 March 2026
AI Meets Strategy: A Product Manager’s Guide to Leading Innovation presents itself as a guide for product managers navigating the growing influence of artificial intelligence (AI). In practice, the book speaks to a much broader audience. As a senior software engineer working closely with product, data, and leadership teams, I found the book relevant to anyone involved in building, scaling, or governing AI‑driven systems in real organizations.
What distinguishes this book from many AI‑for‑business titles is its refusal to treat AI as either a purely technical problem or a strategic one. Instead, the authors convincingly argue that AI success emerges from the interaction between technology, data, organizational design, and communication. That framing aligns well with the lived reality of large engineering organizations and makes the book particularly appropriate for an IEEE Professional Communication audience.
The authors structured the book into four major parts, progressing from foundational concepts to operational practice and real‑world case studies. Early chapters establish why AI changes the nature of product leadership, drawing on cognitive science, data economics, and decision theory. These sections may feel abstract at first, but they serve an important role: they make explicit assumptions that often remain implicit in engineering and product discussions.
The middle sections focus on data platforms, operating models, and machine‑learning delivery. Here, the book is at its strongest. The authors treat data as a product, emphasizing ownership, contracts, quality, and governance. The End-to-End Data for Growth and Enablement (EDGE) data operating model reflects a practical understanding of how data systems fail in practice—not due to algorithms, but due to unclear accountability and misaligned incentives.
The final sections synthesize strategy, ethics, and execution through maturity models, checklists, and industry case studies. While dense, these chapters provide a shared vocabulary that can help cross‑functional teams reason more clearly and consistently about AI initiatives.
The central contribution of AI Meets Strategy is its system-level view of AI adoption. Rather than focusing on model performance or tooling alone, the book emphasizes:
- Clarity of purpose: AI initiatives should be driven by well‑articulated business and user outcomes, not technological novelty.
- Data as infrastructure and communication: Data contracts, semantic layers, and governance models function as communication mechanisms as much as technical ones.
- Operational realism: The book is candid about why proofs‑of‑concept fail to scale and why many AI programs stall after initial success.
- Responsible AI: Ethical considerations are woven throughout the lifecycle, not treated as an afterthought.
From an engineering perspective, this framing resonates strongly. Many AI failures I’ve encountered stem not from technical limitations, but from unclear problem definitions, poor communication across roles, and unexamined assumptions about data and decision‑making.
Although not explicitly marketed as a professional communication text, AI Meets Strategy is fundamentally about professional communication in complex technical systems. The authors repeatedly highlight the need for shared language, explicit assumptions, and traceable decisions across engineering, product, legal, and leadership teams.
Concepts such as data ownership, governance frameworks, and AI maturity models serve a communicative function: they make invisible dependencies visible. The book implicitly argues that AI literacy is less about everyone understanding algorithms, and more about enabling organizations to reason clearly about uncertainty, risk, and trade‑offs.
The book’s ambition is also its main limitation. The volume of frameworks, diagrams, and checklists can feel overwhelming, particularly for experienced practitioners who already operate with informal versions of these models. At times, the material could benefit from stronger prioritization or clearer guidance on tailoring the frameworks to organizational context.
Additionally, while the treatment of responsible AI is thoughtful and practical, it remains procedural. The book explains what organizations should do but spends less time engaging with the deeper ethical tensions and value conflicts that arise in real deployments. For an academic audience, this may feel incomplete; for practitioners, it is sufficient.
These critiques, however, do not significantly diminish the book’s value. The authors are clearly optimizing for usefulness in real organizations rather than theoretical completeness.
AI Meets Strategy succeeds because it reflects how AI work unfolds in practice. It acknowledges that technology alone is not the bottleneck, but communication, organizational structure, and shared understanding are.
For senior engineers, the book provides something rare: a structured way to explain engineering realities to non‑technical stakeholders without oversimplifying them. It does not replace deep technical literature, but it complements it by addressing the organizational and communicative layer where many AI initiatives succeed or fail.
This book offers a concrete, industry‑grounded perspective on how communication shapes the design, governance, and impact of complex technical systems. It is a valuable reference for anyone concerned with making AI not just powerful, but understandable, trustworthy, and effective at scale.
The Emotional Intelligence Advantage: Mastering Change and Difficult Conversations
Amy Jacobson. Melbourne, Australia: John Wiley & Sons, Australia, Ltd. 2025. 240 pages.
Index Terms — dealing with change; emotions; management style
Reviewed by Ginny Redish, IEEE Goldsmith Award winner and a consultant in plain language and usability (ginny@redish.net).
Review posted 4 Feb 2026
As Amy Jacobson writes in her chatty, conversational book, The Emotional Intelligence Advantage: Mastering Change and Difficult Conversation, “Whenever there are human beings involved in any situation, there are also emotions” (p. xvi). However, knowing what emotion you or the people you interact with are feeling at a given moment in a given situation may not come naturally. Jacobson cites research that only 36% of people accurately identify emotions as they occur.
So, Jacobson teaches Emotional Intelligence (EI)—a set of concepts and a specific process for helping us hone our “ability to recognise these emotions, manage the emotions as they play out, and understand the impact they have on the people around us and the outcome of the situation” (p. xvi).
This book is a “how-to” with a specific process for recognizing and working through emotional reactions. Although the primary audience for Jacobson’s book is “leaders” (managers and executives), everyone will benefit from learning about and practicing the steps in the EI process.
In Chapter I, Jacobson describes the EI process using three examples: system changes, a new job, and reading her book. While her first example—system changes—could be a workplace example where executives mandate a change for workers, it could also apply to a personal example like adapting to an unwanted software update.
Jacobson’s EI process has 5 steps:
- Own it: The situation is happening. It is your new reality. Become self-aware. Understand yourself —your past, your values, and how you react to situations.
- Face it: What emotions are you feeling? (There’s no such thing as a bad emotion. The question is whether the emotion is appropriate in this situation.) For a workplace change, these first two steps may focus on the leaders who must understand themselves and their emotions before working with others.
- Feel it: Consider what other people who must deal with this change are feeling.
- Ask it: Listen to those people. Let them ask questions. Work through answers together.
- Drive it: Resolve the situation, which often means accepting the change—but now with emotions externalized, acknowledged, and discussed.
In The Emotional Intelligence Advantage, Jacobson has many useful insights about change and emotions. For example, she points out that change always means losing something. Even when you welcome the change and are happy about it, you have lost something. So, it is important to acknowledge that loss and the feelings of loss before moving forward in the new situation.
For those of us who are technical and practical, understanding and dealing with emotions, especially in the workplace, may feel strange. But emotions are there. And that leads Jacobson to remind us, “It is okay to feel emotions. It is not okay to ignore or downplay them” (p. 83).
Knowing that the EI concepts and process will be new to many people, Jacobson has specific advice, even possible wording, for how to have the difficult conversations that arise as people deal with change in policies, in procedures, in staffing, in assignments, in the tools they must use for their work.
She shares stories of organizations where change went smoothly because leaders and workers went carefully through the entire 5-step process and stories of less successful organizations that ignored or missed steps. You may want to think about your own experiences in different situations, considering how your organization (or you, yourself) used something like Jacobson’s EI process or could have benefited by doing so.
The Emotional Intelligence Advantage is both easy to read and deep in its ideas. I particularly like that Jacobson has worked to make the book interactive. She intersperses her teaching with shaded sections labeled Ask, where she speaks directly to us with questions to consider.
For example, early on, she asks: “How well do you know the people you work with (especially those you lead)?…What can you work on today to better understand how other people are feeling?” (p. 18).
Later, as she talks about the excuses (justifications, limitations) we use to avoid difficult conversations or tasks, Jacobson asks us to consider, “What is ultimately stopping me from doing this? What is the potential outcome that I struggle with the most should it happen? What will I do if this does happen?” (p. 128).
Combining Jacobson’s teaching and insights with using her Ask sections as a workbook can help us all improve our emotional intelligence.
Creativity in PR and Communications: Concepts and Practices for Innovation
Jon Cope and Stuart Mayell. New York, NY: Kogan Page Limited. 2026. 374 pages, including indexes.
Index Terms — Creative processes, creativity in PR, ethics and AI, innovative practices, professional applications
Reviewed by Mickella Rast (mickellarast@gmail.com).
Review posted 4 Feb 2026
Creativity in PR and Communications: Concepts and practices for innovation is a text that is upfront about what it presents and delivers in every way. Jon Cope and Stuart Mayell discuss clear creative concepts, a plethora of creative and innovative practices, and propose a multitude of applications across public relations (PR), communications, and other relevant fields.
The authors break the book into three sections: (1) Introduction to creativity; (2) Creativity in practice; and (3) Issues in creativity. The subject matter in each section remains true to its name. All chapters open with a succinct outline and, with the exception of Section One, Chapter 01: What is creativity?, all chapters close with a breakdown of concluding thoughts, questions, exercises, glossary, and notes. The authors also provide two appendices for further reading and assistance: a creative thinking checklist and an inventory of creative stimuli.
Section One starts with a brief overview of what the book is about, who it’s for, and how it’s organized. The second chapter contextualizes creativity and lays the foundation for the rest of the book. It covers relevant creativity theory, history, and definitions. Spoiler alert: Cope and Mayell do not provide a single definition for “creativity,” but instead review several definitions and famous quotes that readers could find useful. By introducing creativity from several angles, readers can apply whichever approach(es) work best for their personal processes and professional needs.
The remainder of Section One argues about why creativity is necessary in PR and communications and explains how audiences can position themselves creatively. “Creativity” and “innovation” are terms that are vague and may be difficult to reconcile with professional settings, given that they are based in imagination and, often, a lack of precedence or reference. But Cope and Mayell clearly identify how creativity has been tied with PR and communications since the advent of those fields. The heart of this section focuses on the argument that “creativity isn’t just for artists or specialists — it’s a mindset and a skill that belongs at the heart of modern PR practice” (p. 45).
Section Two is a practical walkthrough of the steps Cope and Mayell identify for implementing creativity into your work: research, brief and plan, acquire insight, generate ideas, and evaluate. Section Three is a much-needed and frank address of the current issues with creativity, which are likely to continue into the future: artificial intelligence (AI), ethics, and global communications. Cope and Mayell outline how creativity can be a powerful force, whether for effecting legal change, changing public opinion, spreading knowledge, generating profit, or another goal. At the same time, the authors are upfront about complicating factors like misinformation, political and religious influences, and eroding public trust. Further keeping Creativity in PR and Communications grounded in reality are real-world examples of campaigns, interviews with PR professionals, creative exercises, and tips and thought prompts about your own creative processes.
Notably, the information has a wide range of potential applications, mirroring the overlap between PR and communications and other professional fields that opens Creativity in PR and Communications up to a wide range of audiences. The culmination of these chapters is a final text that is thoughtfully-curated and filled with practical, feasible solutions for integrating creative processes into your work.