AI deepfakes used in every eighth successful fraud scheme in the US

A Gallup survey conducted jointly with the Stop Scams Alliance coalition shows that artificial intelligence and deepfakes are involved in every eighth successful fraud scheme in the United States, with total fraud damage in the country reaching $68 billion over the past year.

How the Survey Was Conducted

The data is based on a survey of 5,173 adult residents of the United States conducted by Gallup in cooperation with the Stop Scams Alliance coalition — an organization that tracks the scale of fraud and develops consumer protection measures. The survey recorded not only the amount of losses but also which tools fraudsters used to convince victims to part with their money.

• Fraud damage in the United States for the year — $68 billion
• 5,173 adult respondents participated in the survey
• Artificial intelligence and deepfakes — in every eighth successful scheme
• Survey organizers — Gallup and the Stop Scams Alliance coalition

How Fraudsters Use AI

Voice cloning and video generation technologies allow the creation of a convincing imitation of a relative's, boss's, or bank employee's voice from just a few seconds of audio, taken for example from social media. A similar scheme involves fake video calls with a deepfake version of a company executive, which have already been used to convince employees to transfer money to fraudster accounts. The growing accessibility of such tools and falling costs are reasons why AI increasingly appears in successful — meaning completed — fraud schemes.

Elderly people and those less familiar with modern AI tools remain particularly vulnerable to such schemes: they are less able to recognize artificially generated voice or video and more frequently accept urgent emotional appeals — requests for help "from" a relative — as genuine.

What This Means

As voice and video deepfakes become cheaper and more convincing, the share of fraud schemes using AI will grow faster than the share of traditional social engineering methods — meaning banks, platforms, and regulators will need to implement voice and video authenticity verification tools in real time more quickly.

Frequently Asked Questions

How Much Money Do Americans Lose Due to Fraud?

According to the Gallup–Stop Scams Alliance survey, fraud damage in the United States over the past year totaled $68 billion.

How Often Is AI Used in Successful Fraud Schemes?

Every eighth successful fraud scheme in the United States involves artificial intelligence or deepfakes, according to a survey of 5,173 adult respondents.