Cristina Criddle / Financial Times:
Sources: OpenAI recently gave staff and third-party groups just days, vs. several months previously, to evaluate risks and performance of its latest models — Testers have raised concerns that its technology is being rushed out without sufficient safeguards