Book Review Series No. 1

Welcome to my new Book Review Series!

Please share your thoughts about the book, and let me know if you have suggestions about non-fiction books to read and review.

Harvard Business Review

I highly recommend HBR’s 10 Must Reads Series, the definitive collection of ideas and best practices for aspiring and experienced leaders selected from Harvard Business Review on topics critical to the success of every manager.

HBR’s 10 Must Reads on AI provides readers with twelve excellent articles on artificial intelligence and machine learning. Based on a recent question from a client about an AI Ethics Committee, that article stood out.

Why You Need an AI Ethics Committee by Reid Blackman (originally published in 2022) explains: “A? strong artificial intelligence ethics committee is an essential tool for identifying and mitigating the risks of a powerful technology that promises great opportunities. Failing to pay careful attention to how you create that committee and how it gets folded into your organization could be devastating to your business’s reputation and, ultimately, its bottom line.”

I agree with Blackman in that I see the need to view and assess the benefits and opportunities of AI through a risk lens to ensure your organization reaps those benefits.

Blackman discusses a 2019 study published in the journal Science which found that AI from Optum prompted medical professionals to pay more attention to white people than to Black people. Health systems used Optum to find high-risk patients who should receive follow-up care. Only 18% of the people identified by Optum’s AI were Black, and 82% were white, which should have been 46% and 53%. Researchers estimated that this AI tool had been applied to at least 100 million patients.

Although the data scientists and executives involved in creating the Optum algorithm did not intend to discriminate, they fell into a shockingly common trap: training AI with data that reflects historical discrimination results in biased outputs.

Since the data used showed that Black people received fewer healthcare resources, the algorithm erroneously inferred that they needed less help.

Notable Ethical Risks

Two of the most notable ethical risks associated with AI are unintended bias and invasions of privacy.

Risks may involve self-driving vehicles that hit pedestrians, AI-generated social media newsfeeds that cause disruption, or risks from reputational, regulatory, financial, or legal threats.

Since AI is built at scale, problems impact all those engaged with it such as all job applicants or mortgage applicants. The potential financial loss is very high.

The article delves into the sources of bias and discrimination and raises some key questions to answer when deciding whether to have an AI ethical risk program. Here are a few:

·?????? Do we face unavoidable ethical risks from AI we design, procure, and deploy?

·?????? Can we mitigate the ethical risks to avoid regulatory fines?

·?????? Can we rebuild consumer trust after what’s viewed as an ethical violation?

Blackman focuses on an AI ethics committee and its members, including ethicists, lawyers, technologists, business strategists, and bias scouts to help answer those questions.

Blackman’s Idea Brief

?? The problem is – bias finds its way into AI and machine-learning models no matter how strong your technology is or how diverse your organization may be.

?? The reason – there are many sources of biased AI, all of which can go undiscovered by data scientists and technologists.

?? The solution – An AI ethics committee can identify and mitigate the ethical risks of proprietary, in-house AI products or procured from a third-party vendor.

AI Ethics Committee

Your internal AI Ethics Committee may be fresh or one that is reassigned to the task.

How will you align your values on ethics with granting authority to the ethics committee? Will you grant them veto power to ensure business impact? Or will you provide oversight from a member of the C-Suite, allowing them to overrule the committee?

AI Ethics Committee Members

Blackman recommends the cross-functional expertise of AI ethics committee members, including Ethics Experts, Lawyers, Business Strategists, Technologists, and Bias Scouts or Subject Matter Experts. The idea is to have the capability to spot biases early in the process.

?My Recommendation

I highly recommend Blackman's article to anyone considering how to proceed with an internal committee surrounding AI integration, be it an AI ethics committee or another AI committee addressing your top values and priorities.

Conclusion

This book review does not cover all of Blackman’s in-depth analysis, for example, the goals you set for AI in your organization, the challenges faced by experts seeking to make predictions, and the quantitative metrics for fairness. I leave that for your reading of this insightful text. ?

?