OpenAI’s decision to scrap the upcoming rollout of a new model may be responsible but could lead us to a flawed view of the power of artificial intelligence, experts have warned.
Overnight, OpenAI said that it was delaying the release of its latest model because it posed a danger if it was released. The system is giving to lying to its users and failing to comply with what humans intended it to do, OpenAI's head of safety systems, Saachi Jain said.
