"We measure AIs to see whether they can pass a bar exam, write working code, and use your computer interface. We test to see how good they are at completing complex tasks, or just impressing humans.
What we donโt have is a rigorous, credible way of evaluating what AI does to us. To our minds, our thoughts, or our communities."
@imrankhan โ who leads our work on psychosocial evaluations of AI โ has a new article on our Substack that explains why we need better AI benchmarks that focus not on raw capability but on how AI is impacting users. Identifying how tech interacts with and changes human nature is the critical first step towards building more humane technology. You can read his full piece here:
https://t.co/zVQWKBdsSV
Thinking of all my friends in the Caucasian Community at this difficult time.
Remember that all the suspicious looks you'll inevitably get over the coming weeks come from a place of ignorance, not hate.
I know these extremists don't represent your proud, peaceful culture โค๏ธ