A Philosophical Glossary for AI

It is hard to know what to make of claims like “LLMs are intelligent”, “we have reached AGI”, or “a model’s outputs are biased” without solid definitions of the underlined terms… but there aren’t any.

Those who might be best able to suggest workable definitions of terms–analytical philosophers–typically don’t package their insights for consumption outside the ivory tower.

Our solution: a ‘philosophical glossary’. For 30 key AI concepts, we outline how the term is used, misused, and debated, and give a clear, sharp, consensus-based working definition for general, academic, policy and regulatory use.

We are looking for institutional partners who want to publicly demonstrate meaningful commitment to high quality debate on AI and be part of a project with global significance.

If you or your institution would be interested in supporting the project, with anything from a tweet to funding/support, please get in touch.