Deepfakes—AI-generated videos, images and audio—are no longer a fringe curiosity. They are a fast-growing threat across the UK, exploited by criminals to scam victims out of money,

impersonate family members and create abusive content that disproportionately targets women and girls. At the same time, hostile actors are deploying deepfakes to spread false information and deliberately mislead the public.

The scale of the problem is accelerating at breakneck speed. In 2025 alone, an estimated eight million deepfakes were shared, up from just half a million two years earlier. That surge has made it harder for people to tell what’s real online, fuelled fraud, and steadily eroded trust in what we see and hear.

A science-led response to a modern threat

To meet this challenge, the Home Office, working alongside the Accelerated Capability Environment, DSIT, DCMS, HMRC and the Alan Turing Institute, has launched a world-first deepfake detection evaluation framework. The initiative brings together technology firms, academic researchers and government specialists with a single goal: rigorously test how well today’s tools can spot harmful deepfakes in real-world conditions.

Rather than focusing on theory, the framework puts detection tools through their paces against everyday risks—impersonation scams, financial fraud and the creation of non-consensual sexual images. By identifying where current technology performs well and where it falls short, the programme aims to raise the bar across the industry. Once completed, the findings will help set consistent standards, guiding companies on how to strengthen their defences against synthetic media.

The government has also backed practical innovation through a Deepfake Detection Challenge hosted by Microsoft. More than 350 experts took part, including teams from law enforcement and the Five Eyes intelligence-sharing community. In a high-pressure, hackathon-style environment, participants were tasked with rapidly distinguishing real media from fake in scenarios tied to national security and public safety.

The exercise tested a wide range of UK and international solutions, sharpening the country’s readiness to respond to deepfake incidents as they happen. It also offered a clear picture of how emerging tools perform when speed and accuracy truly matter.

Setting the global standard

Together, these initiatives place the UK at the forefront of the global response to deepfake threats. By anchoring policy in evidence-based science and creating shared standards, the new framework will help police forces, online platforms and industry partners better identify fraudulent and harmful content. Ultimately, the aim is simple: protect the public from criminal exploitation while restoring confidence in the digital information we rely on every day.

In an era where seeing is no longer believing, the UK’s approach shows how trusted science and cross-sector collaboration can help keep society one step ahead. Photo by mikemacmarketing, Wikimedia commons.