OpenAI Touts New AI Safety Research. Critics Say It’s a Good Step, but Not Enough The company announced a new technique to make the workings of its systems more transparent, but people familiar with OpenAI say more oversight is needed.