The Electoral Commission has launched a pilot scheme to detect and counter political deepfakes ahead of elections in England, Scotland and Wales on 7 May.
The system monitors online content for AI-generated audio and video designed to mislead voters, such as false clips depicting candidates withdrawing from races or making offensive remarks.
It is being delivered in partnership with ACE, the Accelerated Capability Environment, a unit within the Home Office.
Where content raises serious concerns, the Commission can refer material to police and request its removal from social media platforms. All flagged content will be reviewed by a human analyst before any action is taken.
Chief Executive Vijay Rangarajan said a deepfake had yet to meaningfully affect a UK election, adding: ‘We are determined to keep it that way.’
The Commission noted that during the 2024 general election, around a quarter of voters surveyed reported seeing or hearing a deepfake.
The pilot will run until June to allow time for evaluation.
