OpenAI has pledged to regularly communicate instances where its artificial intelligence models deviate from intended human behaviors. This initiative aims to improve transparency regarding AI safety and alignment issues.