The "AI Fraud Accountability Act" establishes criminal and civil prohibitions against using digital impersonations, created with artificial intelligence or other technology, to defraud individuals. It defines digital impersonation as a visual or audio depiction of an identifiable or imaginary individual that is indistinguishable from an authentic or real person. The bill makes it unlawful to falsely pose as such an individual with the intent to defraud, with exceptions for authorized law enforcement and intelligence activities. Violators face significant penalties, including fines, up to three years imprisonment, and forfeiture of any proceeds or property used in the crime. The Federal Trade Commission (FTC) is also empowered to enforce these prohibitions as unfair or deceptive acts, leveraging its existing authority, and the Act establishes extraterritorial jurisdiction for these offenses. To enhance detection and prevention, the bill mandates the Secretary of Commerce, through NIST, to convene a working group of government agencies, private sector industries, and AI experts. This group will develop and publish best practices for recognizing, detecting, preventing, and tracing fraudulent digital impersonations, with public workshops and annual reports to Congress. Finally, the bill addresses the international nature of this fraud by requiring the FTC to identify top foreign countries where violations originate and enter into cooperation agreements. The Attorney General is also tasked with reviewing and modifying international agreements to strengthen assistance in prosecuting extraterritorial digital impersonation fraud that harms U.S. persons.