The government included provisions targeting artificial intelligence-generated electoral disinformation in the Representation of the People Bill, extending the existing electoral law framework to address the growing threat of AI-synthesised video, audio and image content being deployed to mislead voters about candidates, policies or electoral processes. The provisions represent the first specific legislative response to the deepfake threat in the context of UK elections and were supported across party lines despite some technical concerns about implementation.
The core requirement under the new provisions is that AI-generated content used in electoral communications must be labelled as such, allowing voters to identify when they are viewing synthetic media rather than authentic recordings. The labelling obligation falls on the creator and publisher of the content, with enforcement responsibility shared between the Electoral Commission — which has powers over electoral offences — and social media platforms under their existing obligations under the Online Safety Act.
Technical experts noted that effective enforcement of the labelling requirement depends on reliable detection of AI-generated content — a capability that is advancing but not yet infallible. The most sophisticated AI image and video generation techniques can produce outputs that are difficult to distinguish from authentic footage even with forensic analysis, creating a risk that the labelling requirement, while well-intentioned, could be circumvented by bad actors with access to the best available technology.
The Electoral Commission welcomed the legislative clarity but called for additional resources to enable it to respond effectively to potential violations during the May 7 elections. The tight timing between the Bill's passage and polling day meant that full implementation of enforcement systems would need to be accelerated significantly.
