Trustworthy AI: A Business Guide for Navigating Trust and Ethics in AI

While EU legislators and US policymakers such as NIST think about how to define trustworthy AI, the author makes a compelling business case for thinking broadly about how to maximize AI's potential in the eyes of the stakeholders, including customers and employees. Although the way the author weaves in a comprehensive discussion of legal standards that warms the heart of this attorney, first and foremost the book is about business and how to make sure that AI at least meets the expectations of stakeholders, if not delights them. While the author is technically well-grounded, this is not a technical book either in the sense of discussing complicated technology, nor in the sense of suggesting a mechanical approach to achieving trustworthiness. Such an approach would surely fail. Instead, the author, in very accessible language, lays out the factors that senior managers should consider when implementing AI solutions individually and at the enterprise level, factors designed to ensure that people actually embrace and benefit from AI.

Finally. An understandable book on AI ethics. There is a lot of headlines and news circulating about AI ethics in the academy, the industry, and governments. There are wide ranging discussions about competing lists of principles for AI ethics, worries over the potential harms of ethics washing in the AI industry, and seemingly endless arguments about the philosophical details. This book provides an easy way to understand whats real behind the hype and actionable steps anybody can take.

A book that will help socially conscious business leaders navigate the challenges as AI permeates all aspects of an organization. The author simplifies the problem by using illustrations that also add a touch of humor to the subject. This book can come in handy while dealing with AI issues not only in the real world but also in the metaverse.

Love this book! Throughout, Beena takes the concepts of Trustworthy AI out of the theoretical and into the practical. She carefully defines in detail each principle of Trustworthy AI, and then brings it to life with real world applications. A must read for executives.

A good book on AI from a different perspective. It could be a bit more technical but still a good read.

One of the few books that explains AI ethics with real depth for more than just IT experts. It goes deeper than the theoretical, high-level looks at AI ethics. And the book helps ground the discussion so you can take tactical steps toward making trusted AI real and tangible. It’s a starting point for asking the right business questions and a guide for the journey to operationalize AI ethics and trust. Highly recommend it for anyone considering or undergoing technology transformation.

As businesses across the globe transition towards AI and the potential for growth that it represents, the need for a pragmatic guide to the ethical use of AI has increased exponentially. Enter "Trustworthy AI", which offers a direct outline for the implementation of critical strategies in the space of trust and ethics in AI. I'd highly recommend getting your hands on this book!

Disclaimer: I did not pay for this book, it was given to me but no promises were given for a favorable review, which has probably worked out for all parties concerned. The back of the dust cover says it is an 'actionable guidebook on AI ethics for executives, technologists, ethicists, and users." I disagree on two points there, it is squarely positioned at executives who need to know just enough to sound like they know, and I found very little actionable. But more on that below. The 1st chapter "A Primer on Modern AI" is actually pretty well written (although, not attempting a better definition in terms of AI itself, seems central to the book. Just to say ' it is a lot of things ' is a bit of a cheat. Two minor issues to get out of the way. -= Minor Gripe #1 =- "In the 1930s, [Alan] Turing demonstrated that mathematically that rules-based code could solve algorithmic problems...." - this is probably referring to his paper 'On Computable Numbers' , the best book that subject would be this one, available on Amazon -> https://www.amazon.com/gp/product/0470229055/ this is not an accurate portrayal of his paper. But the this book continues "...and it was he who developed the eponymous testing for interrogating the presence of machine intelligence". No. This is just incorrect. I would not diminish any of his work, the man was a genius and sadly persecuted in his own time. But Turing wrote 'Do Machines Think?' and posited if a human judge could not differentiate between a human player and a computer in what he described as 'the imitation game' then readers could cede that the computer could be said to 'think'. He did not make a framework for testing natural language, or any thing like that. The term 'Turing Test' is used more out of respect of his early work and less an reference implementation of his work. -= Minor Gripe #2 =- Reading through this, I can't help but wonder if if this was a single author - or if there were some other contributions through editing (or publisher inclusions?). So much of the page count is inflated with * pointless quotes on single page. "To every action there is always opposed an equal reaction - Isaac Newton. Agree, but Newton was writing about physical bodies. * A fictional use case (BAM Industries) which doesn't really go anywhere. Maybe in concept this book was going to go the 'Phoenix Project' direction? But it didn't. The use case gets mention periodically but only to suggest large companies have people who do different roles. Got it. Would have been better to talk to actual companies and get real use cases. * Cartoons, some of them printed sideways, always with word balloons and Comic Sans text that look they were done in MS Word - added nothing to the book. -= Major Complaint(s) =- While some concepts are described there isn't really anything actionable suggested other than to acknowledge that it will require people, process and technology. That isn't actionable. There is almost nothing tangible to anchor this to suggest this is a 'business guide for navigating trust and ethics in AI' as the byline suggests. The few items mentioned (I'm paraphrasing) would be 'anonymize data for PII', 'have principles documented' and 'institute an AI ethics review board' - all good suggestions but hardly enough to hang an entire book upon. Secondly, in describing bias the author (ironically?) injects their own. We can talk about the desire for equal treatment - but without getting into sensitive specifics, when the raw data is pretty clear where the outliers are with regard to how crime is distributed -- there are no doubt many factors at play. Urban areas have more people, and crime is a function of proximity (it is no surprise very remote areas have lower crime rates - some people in wild Alaska may go days without seeing anyone outside their own family. So not a lot of car-jackings going on there). In a book promoting the correct and careful application of AI it is irresponsible to let the 'woke talking points' cloud matters of fact. tldr; Lots of filler, there is better and more concise material for free on the web.