OpenAI said Friday that it fired three safety researchers for a 'breach of trust,' defending the dismissals after the trio accused the company of putting interests before safety.