My ‘AI Truth and Transparency Act’ would require durable disclosure when realistic synthetic media is presented publicly. Stronger penalties would apply when AI is used for fraud, impersonation, fabricated emergencies, market manipulation, or malicious election interference.

Americans should also possess enforceable protections against unauthorized commercial or fraudulent replicas of their face and voice, while preserving legitimate exceptions for parody, criticism, and public commentary.

AI regulation should be based on risk, like a risk-tiered system. Ordinary creative tools should not face the same requirements as systems controlling medicine, transportation, infrastructure, finance, law enforcement, or weapons…

The federal government would also be prohibited at any time from establishing a social-scoring system based on lawful speech, beliefs, purchases, associations, or political behavior via a ban by my administration.

My policies:

• Pass the AI Truth and Transparency Act.

• Require disclosure of realistic synthetic media.

• Protect personal faces and voices from fraudulent replication.

• Preserve parody, journalism, and legitimate commentary.

• Regulate AI according to its level of risk.

• Require stronger testing for high-impact systems.

• Guarantee human review of major automated government decisions.

• Ban federal social scoring.