The Seagal images look like they are from Dall-E v2. OpenAI has only given selective API access to Dall-E v2. There are a number of safeguards in the model itself for misinformation. There are business reasons not to give the API away for free, which I think are the main reasons they haven't released it yet. Also last time they claimed misinformation issue to limit a model release was GPT-2 and that was so they could make the news instead of real concerns. They have been kind of a dirty company lately, at least very contradictory to their initial aims.
Google just created their own comparable model as well.