@philosophyminis: In the 1980s, Frank Jackson gave us one of the most memorable thought experiments in philosophy. When Jackson was writing, he was writing about human thoughts and human knowledge. But it's also a great way to introduce a problem in the #philosophy of #artificialintelligence Here's how the story goes: Imagine there is someone called Mary who is in a black-and-white room with no contact with the outside world. Everything is black and white. But she has access to the vastness of Wikipedia. Over her years, she spends every day and every minute learning about the colour red. She learns that blood is red, roses are red, and that #red is associated with #Love , #passion , danger, or luck. But she has never seen red. Now, one day, Mary leaves her room, and she sees a vibrant red strawberry. She has an experience of red. The question that Jackson asks is: Does Mary learn anything when she sees red? Of course, today, AI plays the role of Mary. Large language models are fed billions and billions of words about billions of things. AI can tell you ten things that are red. It can tell you what happens in the cones in your eyes when they pick up light waves to send to your brain, to process it into the feeling of red. It's processed images of red and can reproduce red in pictures. And it recognizes colours. It picks out pigments, and it can isolate shades. But the question that Jackson asks is, "Does #AI understand what red is? Does it feel redness, or does it feel anything at all?"

Jonny Thomson

Open In TikTok:

Region: GB

Monday 05 August 2024 13:26:29 GMT