OpenAI flags 'unexpected and concerning' AI behaviour. Can it be avoided?

OpenAI flags 'unexpected and concerning' AI behaviour. Can it be avoided? Meetings & Conventions Asia