OpenAI updated its safety framework—but no longer sees mass manipulation and disinformation as a critical risk | Fortune