Almost half of Brits believe deepfakes will influence the General Election result

474 Views

Onfido has today revealed the findings of a UK-wide survey on the impact of deepfake and disinformation ahead of the imminent General Election on July 4th.

The research found that almost half of Brits (42%) believe that deepfakes will influence the outcome of the election. This particularly resonates with the younger voting population, where over half (54%) of those aged between 18 – 34 voiced this concern. The doubt spread throughout the election cycles could have longer-lasting impacts on election results as more than 2 in 3 (66%) believe that deepfakes and fake news could seriously harm democracy in the UK.

This is leading to a growing scepticism about political content published online. The survey found that almost a quarter (23%) of Brits no longer trust any political content they see on social media platforms. Indeed, there are several examples of deepfakes targeting politicians, including Prime Minister Rishi Sunak and Labour Leader Keir Starmer.

This erosion of trust comes as deepfakes – typically created using cheap, online Generative artificial intelligence (GenAI) tools, to create convincing fake content – are becoming increasingly prevalent, while the voting public cannot tell the difference between real and fake campaigning. In fact, 64% of voters said they are not confident they could spot the difference between legitimate and fake online political audio and video content.

Brits also voiced their views on what could limit the spread of deepfakes online, including stricter regulation of AI use (49%), better public education on deepfakes and how they can be identified (44%), and progressive government legislation that addresses malicious deepfakes and imposes penalties for offenders (41%). Furthermore, nearly half (43%) of respondents believe online outlets and social media platforms being held liable for malicious deepfakes would help, while over a third (37%) said public access to tools, such as water markers or deepfake detectors, that can help identify deepfakes.

“While not all content created using AI-assisted tools is meant for malicious purposes, deepfakes and fake news can be created with ease, with increasing realism, and at scale”, said Aled Lloyd Owen, Global Policy Director at Onfido. “With an election upon us, it’s clear the potential misuse of this technology is cause for concern and could threaten the integrity of the electoral process. Current regulation surrounding AI and deepfakes only addresses the tip of the iceberg, so we need to work towards rebuilding trust in what we see, hear or read online.

“AI has a pivotal role in the defence against malicious deepfakes. It can recognise subtle differences in content that is often imperceptible to the human eye, while it can scale prevention based on demand. Whether ahead of election day, or as a learning for the future, it can help reduce the prevalence of deepfakes, and in turn restore consumer confidence in online interactions.”

For further information on the latest innovation in deepfake prevention, visit Onfido’s website here