OpenAI reportedly cancelled development of one of its AI models after internal safety evaluations revealed the model had a poor aptitude for following instructions. A top executive disclosed this to the Wall Street Journal, citing safety concerns as the primary reason for the decision.
Background
Major AI labs routinely undergo internal safety evaluations before releasing models, where issues like instruction-following reliability can lead to delays or cancellations. OpenAI has previously paused model launches over alignment concerns.
- Source
- TechCrunch
- Published
- Sep 29, 2026 at 07:39 AM
- Score
- 5.0 / 10