OpenAI’s ironic credibility problem (www.understandingai.org)

How do we know the company’s stated reasons for firing three safety researchers are its actual reasons?