Deepfakes (fake videos, images or audio created using artificial intelligence) are a rapidly evolving threat to the UK. Criminals are using them to trick people into handing over money, to impersonate loved ones, and to create harmful and abusive content, especially targeting women and girls. Meanwhile, hostile actors are using them to create false and inaccurate information to deliberately mislead the general public.
Its scale is rising rapidly. In 2025, around eight million deepfakes were shared, compared with just half a million two years earlier. This huge rise has made it challenging for people to know what online content is real, increased the risk of scams, and damaged trust in what we see and hear.
A science-led response
To tackle this threat, the Home Office in collaboration with the Accelerated Capability Environment, DSIT, DCMS, HMRC and the Alan Turing Institute launched a world‑first deepfake detection evaluation framework. It brings together experts from technology companies, academics, and government specialists. The task to evaluate the performance of detection tools to identify harmful deepfakes across different applications.
These tools are tested against real‑life problems such as impersonation, fraud, and non‑consensual sexual images to see where current technology works well and where it needs to improve. Once complete, the results will set new standards for industry, guiding companies on how to better detect and prevent deepfakes.
The UK Government also recently funded a Deepfake Detection Challenge, hosted by Microsoft. More than 350 experts were involved, including teams from law enforcement and the Five Eyes community – an intelligence-sharing alliance comprised of several nations. They faced high‑pressure, time-sensitive tasks to identify the difference between real and fake media linked to national security and public safety. This hackathon- style event helped assess a range of UK and international based solution-providers and is another leap towards strengthening the UK’s overall ability to deal with synthetic media.
This work has put the UK at the forefront of the global response to deepfake threats. By creating consistent standards, the framework will help police, industry and online platforms better identify fraudulent and harmful content and will ultimately, keep the UK general public safe from criminal exploitation. It also showcases the vital role that trusted, evidence-based science advice plays in protecting society.