OpenAI unveiled new research outlining the organization’s approach to make certain AI reasoning styles continue to be aligned with the values of their human builders. The startup used “deliberative alignment” to generate o1 and o3 “Believe” about OpenAI’s security policy.OpenAI is blaming among the list of longest outages in its histori