The move comes after a 16-fold increase in online deepfakes in the past two years

Britain has launched a world-first framework to detect and combat deepfakes after more than eight million such AI-generated videos, images and audio files were shared online in 2025, up from 500,000 two years earlier.

Criminals are increasingly using deepfakes to defraud victims, impersonate loved ones, and produce harmful content targeting women and girls. 

Hostile actors are also deploying the technology to spread deliberate misinformation among the general public.

In response, the Home Office, working alongside the Accelerated Capability Environment, DSIT, DCMS, HMRC and the Alan Turing Institute, has developed a deepfake detection evaluation framework.

UK at the forefront of the global response to synthetic media threats

The scheme brings together technology companies, academics and government specialists to assess detection tools across a range of harmful applications, including fraud, impersonation and non-consensual sexual imagery.

A Home Office spokesperson said: “This framework will help police, industry and online platforms better identify fraudulent and harmful content and ultimately keep the public safe from criminal exploitation.”

The government also funded a Deepfake Detection Challenge hosted by Microsoft, which drew more than 350 experts, including representatives from law enforcement and the Five Eyes intelligence alliance. Participants faced time-sensitive tasks identifying fake media linked to national security and public safety concerns.

Officials said the initiative has placed the UK at the forefront of the global response to synthetic media threats, with the framework set to establish new industry standards for deepfake detection and prevention.​​​​​​​​​​​​​​​​