Provider policy / Child safety · Governance & regulation · Safety & wellbeing
The draft Humanist AI Code of Conduct says models should support children’s development and human relationships, avoiding dependency and substitution for trusted people.
Why it matters
It gives designers and evaluators a concrete example of intended relational boundaries, beyond prohibited-content rules.
Read with care
A provider’s proposed behaviour standard is not independent evidence that its systems consistently meet that standard. The document was published for consultation.
What was checked
Draft status, publication date and child-safety section checked.