Microsoft and the UK will collaborate to create a system to automatically detect “deepfakes” in the online environment

The British government announced on Thursday that it will work with Microsoft, academia and experts to develop a system to identify “deepfake” online material as it seeks to set standards to combat harmful and misleading content generated by artificial intelligence (AI), Reuters and Yahoo News reported.

Although manipulative and deceptive material has been circulating on the Internet for decades, the rapid adoption of generative artificial intelligence tools has heightened concerns about the scale and realism of deepfakes. These are usually considered images, videos or audio recordings in which a person's face, voice or identity is replaced or imitated very realistically so that they appear authentic, even though they are completely fabricated.

Britain, which recently criminalized the creation of intimate images without consent, said it was working on an evaluation framework for detecting “deepfakes” aimed at setting consistent standards for evaluating detection tools and technologies.

“Deepfakes are being weaponized by criminals to deceive the public, exploit women and girls and undermine trust in what we see and hear,” Liz Kendall, London's technology minister, said in a statement.

Explosive increase in the number of “deepfakes” circulating on the Internet

Prime Minister Sir Keir Starner's government said the framework will assess how technology can be used to analyse, understand and detect harmful deepfake material, regardless of its source – by testing deepfake detection technologies against real-world threats such as sexual abuse, fraud and identity theft.

The UK government also said this would help it, in collaboration with law enforcement, gain a better understanding of areas where gaps in detection persist. The London cabinet added that the framework will be used to set clear expectations for industries around standards for detecting deepfakes.

According to government figures, around 8 million deepfakes were distributed in the UK last year, rising from 500,000 in 2023.

Governments and regulators around the world, struggling to keep up with the rapid evolution of AI technology, were spurred into action this year after Elon Musk's Grok chatbot was found to be generating sexualized images of real people, including children, without consent.

The UK's communications regulator and data protection authority are conducting parallel investigations into the Grok chatbot.