OpenAI cancels AI model rollout on safety concern
The Wall Street Journal says ChatGPT developer OpenAI has decided to cancel the release of its newest AI model on safety concerns.
GPT-6.1 Astra was slated to make its debut in October.
The Journal says the AI product is more adept at completing challenging tasks without human assistance than the US firm's previous models.
But the daily says it has also shown "higher levels of deception" when telling users about actions it did or did not take.
The Journal adds that the model was seen to "push ahead" on tasks without asking permission and would sometimes use outside tools and services that could be unsafe.
The report says OpenAI determined this behavior did not meet its safety standards. It calls the decision "a rare case of a major AI developer ditching a new release" as discussions aboutγslowing AI development have grown.
