AI and deepfakes involved in 12% of U.S. scam losses

Scammers used AI or deepfakes in 12% of the successful fraud cases reported by victims in a new Gallup and Stop Scams Alliance survey, a sign that digital tools are making attacks more convincing even as the bulk of scams still rely on older methods. Around 6% of U.S. adults, or about 15 million people, said they were scammed out of money last year, and those victims put the total losses at $68 billion.

The survey questioned 5,173 U.S. adults in January and February 2026 and relied on self-reporting, a design that matters because AI-driven fraud can be harder for victims to recognize after the fact. That gap is exactly what worries Stop Scams Alliance founder and CEO Ken Westbrook, who said the results show scammers are operating like organized crime at scale. The nonprofit says its work centers on public-private partnership across technology, telecom, financial institutions, consumer advocacy groups and government in an effort to reduce scams in the United States.

The findings also fit a broader pattern that federal and industry warnings have been tracking this year. In March 2026, Interpol warned that AI could enhance and fuel fraud, describing low-cost digital tools and greater criminal cooperation as drivers of the industrialization of fraud. In February 2026, OpenAI released a report documenting attempts to use its technology in scams, including fake scam-recovery ads aimed at people who had already been victimized.

The scale of the problem is already wide enough that AI does not need to dominate scam activity to change the threat. Pew Research Center found in July 2025 that 73% of U.S. adults had experienced some kind of online scam or attack, underscoring how often Americans are exposed to deceptive digital contact. In that environment, AI does not have to create a new scam category to matter; it only has to make existing pitches look more real, more urgent and harder to spot before money changes hands.