Welcome to December 2024’s “Book of the Month” column. This month, we’re featuring “AI Governance Comprehensive: Tools, Vendors, Controls, and Regulations” by Sunil Soares, available for free download on the YourDataConnect (YDC) website.
This book offers readers a strong foundation in AI governance. While the emergence of generative AI (GenAI) has brought AI governance to the forefront of discussions, Sunil emphasizes the extensive nature of various other AI use cases and the necessary governance for each. Whether the application for AI governance is machine learning, GenAI, or another use case, this book proves to be an indispensable reference. “AI Governance Comprehensive” provides detailed information on use cases, regulations, controls, and the tools that vendors are providing in the marketplace today. Additionally, the book provides a framework for AI governance, a 13-step wheel that describes the activities of AI governance, complete with the roles, responsibilities, and activities in each step of the governance process.
One of the key themes in “AI Governance Comprehensive” is the mitigation of bias in AI systems and the crucial role of human oversight. AI systems, while powerful, are not immune to biases that can arise from the data they are trained on or the algorithms they employ. These biases can lead to unfair or discriminatory outcomes, which can have significant ethical and societal implications. To address this, the book emphasizes the importance of maintaining a human-in-the-loop approach. This involves integrating human judgment and decision-making at critical points in the AI lifecycle to ensure accountability, transparency, and fairness. By doing so, organizations can better manage and mitigate biases, ultimately fostering more equitable and trustworthy AI systems. The presence of human oversight also enables continuous monitoring and adjustment, ensuring that AI systems align with ethical standards and societal values.
One of the book’s strengths is its practical applicability, demonstrated by tying specific industry use cases to relevant regulations and activities. For example, one of the use cases the book explores is using AI as part of a healthcare system. This particular use case highlighted the creation of an AI council to address specific challenges faced by the organization. The use case goes into detail around specific formal roles and key documentation points, such as the summary of what the model does, the risk associated with the model, the owner, how it was trained, and what type of model it is. The book is filled with use cases like this, complete with the regulations that govern it, and often references to the tools used to assist with the governance operations. To complete the authenticity of these use cases, the book cites the specific regulations, tools, and real-world examples for the reader.
Sunil spends some time discussing the transparency of AI models and what to watch for. A feature described in the book is explainable AI (XAI), which emphasizes the importance of understanding how a model achieves its results. The book also touches on watermarking AI results to prove that content was AI-generated. Additionally, it discusses content creators using tools for data poisoning, which can cause models that use content without consent to learn unpredictable behaviors.
An important note is that the book provides extensive detail on vendor solutions available to assist organizations with AI governance. It quickly becomes an invaluable reference for readers, showing not only how specific tools are used, supported by use cases in the book, but also highlighting the specific areas where different vendors excel and how to maximize what is available in the current marketplace for AI governance.
In conclusion, “AI Governance Comprehensive” stands as an essential reference for anyone involved in AI governance today. The book’s detailed exploration of use cases, regulations, and tools provides readers with practical insights and actionable strategies. By addressing the complexities in AI governance, it equips professionals with the knowledge needed to navigate the evolving landscape of AI. Furthermore, the inclusion of real-world examples and vendor solutions makes this book an invaluable guide for implementing effective AI governance frameworks. For those seeking to understand and implement AI governance, this book is a must-have resource.
More About the Author
Sunil Soares is the founder and CEO of YourDataConnect (YDC), focused on AI governance. Prior to this role, Sunil was the founder and CEO of Information Asset, a data management firm, which he sold to private equity.
Sunil is the author of 12 books on data management and AI governance, including The IBM Data Governance Unified Process, Selling Information Governance to the Business, Big Data Governance, Data Governance Tools, Data Governance Guide for BCBS 239 and DFAST Compliance, The Chief Data Officer Handbook for Data Governance, and AI Governance.
In the past, Sunil also worked as an auditor at PwC and as a management consultant at Booz and Company. Sunil was a member of the Institute of Chartered Accountants of India and has an MBA in Finance from the University of Chicago Booth School of Business. Sunil also holds the Artificial Intelligence Governance Practitioner (AIGP) certification from the International Association of Privacy Professionals. He has also successfully completed the IEEE CertifAIed™ Assessor Training for AI ethics assessments.