There is currently a lot of talk a lot about AI ethics; the topic is large and multifacted and the current discourse is fragmented. Ethical issues are intristically intertwined with similar challenges around information governance and trust. And although "AI" is used to describe a large category of machine-enabled automation, the rapid rise and adoption of Generative AI technologies has generated some urgency around ethical considerations.

Such ethical considerations around AI include or intersect with issues of accuracy, misinformation, environmental concerns, bias, and reputational risk (among others). Is there a way to try to make sense of all of these issues? In business and societal contexts, what are the risks involved? Can we use existing ethical frameworks to help us grapple with ethics in AI?

While other ethical frameworks (like deontology) consider ethical questions to be decided according to a set of principles (and regardless of outcome)--an relevent example here  could be "automatiion is good"--consequentialism (as the name implies) considers the outcomes of actions to be the determining factor when making ethical judgements.

Narrowing down the discussion (from the world of AI ethics writ large) to the question :"what are the ethical risks involved in deploying GenAI at scale?" and drawing on the framework of consequentialist ethics allows a more focused examimation of the risks and implications involved.

In this talk I will:

• Briefly examine a few ethical frameworks
• Propose consequentialist ethics as a good fit for the discussion of GenAI
• Examine ethical risks involved in deployment at scale
• Suggest ethical and low-risk ways forward

Speaker: Robert Kasenchak

Bob Kasenchak is an information architect at Factor. As a taxonomist and ontologist with an interest in knowledge graphs and Linked Data, he has worked for over a decade building and implementing taxonomy projects for publishing, enterprise, technology, and e-commerce clients. He brings experience with information modeling and semantic software to client-focused metadata and vocabulary projects. Bob holds an MM in Theoretical Studies from the New England Conservatory of Music and a BA in Liberal Arts from St. John’s College, Santa Fe; he put in 5 years towards a PhD in Music Theory at the University of Texas before abandoning academia for the information industry. A frequent writer and presenter on semantic topics in conferences and journals, Bob’s current research interests include ontologies, knowledge graphs, and text classification.