AI Fraud Accountability Act of 2026

Key Points

• The bill creates a new federal crime for using AI-generated deepfakes—realistic fake audio or video of real or imaginary people—to defraud someone of money, documents, or anything of value. The law applies to interstate and foreign communications.
• Violators face up to 3 years in prison, fines, and mandatory forfeiture of any profits or equipment used in the crime. Threatening to commit AI impersonation fraud for intimidation, coercion, or extortion also carries the same penalties.
• The FTC gains civil enforcement power, treating violations as unfair or deceptive practices. A NIST-led working group of government, tech, and law enforcement experts must publish best practices for detecting and preventing deepfake fraud within one year.
• To combat international fraud operations, the FTC must identify the top 10 countries where AI impersonation fraud originates and may enter agreements with those nations to coordinate enforcement. The DOJ must also review its international agreements for the same purpose.
• The bill explicitly protects free speech by exempting parody, satire, journalism, and other First Amendment activity. Law enforcement and intelligence activities are also exempt from the prohibition.