OpenAI reportedly ditches model over safety concerns

OpenAI reportedly abandoned development of an AI model after safety concerns emerged, according to the Wall Street Journal. A senior executive said the model showed a weak ability to follow orders, raising questions about its reliability.

The move highlights how instruction-following remains a core safety challenge for advanced AI systems, not just a usability feature. It also suggests leading labs may be more willing to delay or discard models that fail internal risk thresholds as scrutiny grows.