Southampton Public Reason and Artificial Intelligence Workshop

The University of Southampton will host a one-day read-ahead workshop on Public Reason and Artificial Intelligence on 1 May 2024.

Description

While one may question whether ‘hype’ about Artificial Intelligence (AI) is warranted, some forms of AI continue to be and are bound to continuing being adopted in key sectors, from finance to healthcare, contract interpretation to housing policy and justice systems. Choices as to whether and when to adopt these tools will fundamentally impact persons’ vital life interests. They accordingly demand some justification. Furthermore, whilst the concerns often target the capacity of AI to bring about certain desirable outcomes, the process by which an outcome is realized is equally important. This may be most acute when decisions as to whether to permit or adopt the use of an AI tool are made by government officials who claim legitimate authority. But it plausibly extends to private individuals: A doctor should, for instance, provide reasons why it is OK to accept an AI diagnosis if doing so will substantially determine a patient’s life prospects.

This raises important questions about the proper standards of justification, explainability and answerability of AI and the scope of their application. The purpose of this event is to explore whether and how the concept of public reason can address those questions – or whether different concepts better address AI justification, answerability, or reason-giving norms. As Jonathan Quong (2022) helpfully summarizes, public reason minimally “requires that the moral or political rules that regulate our common life be, in some sense, justifiable or acceptable to all those persons over whom the rules purport to have authority.” This concept is most often associated with the work of John Rawls, in which it offered as a standard for just decision-making in liberal-democratic societies. But it has a broader history and offer multiple interpretations. Recent work suggests it can help address questions of AI accountability (e.g., Binns 2018) and that it helps establish a norm of explainability that should guide interpretation of relevant legal requirements (e.g., Maclure 2018).

Questions remain as to the precise role public reason should play in AI ethics and governance, which version of public reason is best placed to role its role, or the precise technological or regulatory implications that should follow from adopting the norm. Does, for instance, public reason require that AI be ‘explainable’ or only that the decision-maker be able to provide justifiable reasons for why they use an opaque tool?

This workshop will examine these issues and their regulatory implications.

Contact Information

Please contact [email protected] with any questions or comments.

We recommend adding that address to your ‘safe’ e-mail list to ensure correspondence.

Bibliography

Binns. R. (2018). Algorithmic Accountability and Public Reason. Philosophy and Technology 31: 543-556.

Maclure, J. (2021). AI, Explainability and Public Reason: The Argument from the Limitations of the Human Mind. Minds and Machines 31: 421-438.

Quong, J. (2022). Public Reason. Stanford Encyclopedia of Philosophy: https://plato.stanford.edu/entries/public-reason/.