They seem to be adding a policy based on ai-policy.dev/policies/human-responsible/ github.com/colibri-social/colibri.social/…/170 As somebody with somewhat pro-AI views, I think this is quite reasonable.