Study finds AI is not conscious, yet
Via Semafor
| Researchers attempted to see whether AI is conscious, and determined — cautiously — that current models probably aren’t, but future ones could be. The study sought to understand whether AIs can think about what they’re thinking, in the hope that that tells us whether they have real inner life. The research does not solve the millennia-old hard problem of consciousness, the AI-and-philosophy-obsessed writer Scott Alexander noted, but makes useful progress. AI’s rise has, Alexander said, forced researchers into doing “philosophy with a deadline.” If AIs have conscious experience, we may have moral obligations to them, and questions bandied around by philosophers since Plato have real-world implications. Alexander also noted that people already often treat AI as though it were conscious. |
0 Responses
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.