Overview
- The proposal mandates clear disclosure that an AI companion is not human, with recurring notices at least every three hours when the operator knows the user is a child.
- Operators must adopt protocols to prevent content that encourages self-harm or violence and provide crisis resources when users express thoughts of self-harm.
- AI companions would be prohibited from producing sexually explicit visual material for minors or instructing them to engage in sexual conduct.
- Enforcement would rest with the Pennsylvania Attorney General, who could seek civil penalties of up to $10,000 per violation.
- Backers cite child-safety concerns and recent lawsuits as motivation, noting state action continues even as a federal order allows challenges to state AI laws but exempts child-protection measures.