OpenAI has acknowledged an investigation into a social engineering incident where company personnel engaged with Russian hoax callers posing as official contacts to discuss defense applications of artificial intelligence.