Report reveals UK voters' fears over AI & deepfakes in election
As the United Kingdom's general election approaches, a new report indicates growing concerns regarding the influence of artificial intelligence (AI) and deepfakes on the democratic process. According to research conducted by Jumio, 60% of Britons expressed apprehension about the potential for AI and deepfakes to sway election outcomes.
The study, which canvassed the opinions of over 2,000 adults from various regions across the UK, also highlighted that only one-third (33%) believed they could easily identify a deepfake of a political figure. This has brought to light significant worries about the reliability of information being disseminated online as voters prepare to head to the polls on 4 July.
Further findings from the research reveal a shifting dynamic between the British public and online media. Compared to the last general election in 2019, 64% of respondents stated they were more sceptical of political content encountered on the internet. Interestingly, Northern Ireland emerged as the region with the highest level of scepticism, with 79% of participants expressing doubt about the authenticity of online political information.
Fewer changes were noted in the levels of trust in traditional news sources, with 56% of Britons reporting no change in their trust in print or broadcast news. Nonetheless, 25% indicated they trust these traditional news formats less than before. On the other hand, 40% of Britons now trust social media sources less compared to the period preceding the advent of generative AI technologies.
The research also indicated that Britons desire stricter regulation concerning AI. Over half (53%) believe current UK laws on AI are insufficient, with only 26% trusting the government’s capabilities in regulating this rapidly evolving technology. This sentiment underscores a call for more robust measures to safeguard the integrity of electoral processes in the age of artificial intelligence.
Stuart Wells, Jumio's Chief Technology Officer, emphasised the importance of addressing these concerns. "With the UK heading to the polls, it's vital that we have an open conversation about the role that generative AI and deepfakes could play in the national debate," he said. "The public's lack of confidence in their ability to identify fraudulent content online is concerning, and more needs to be done to educate consumers on how to spot deepfaked content, and how to report it should they see it."
Wells also noted the responsibilities of online organisations in combatting the rise of deepfakes. He urged such entities to implement advanced verification systems that could prevent the dissemination of manipulated content. "Online organisations should implement multimodal, biometric-based verification systems or other deepfake detection mechanisms to keep deepfakes from influencing voters in the days leading up to the election and beyond," he said.
The research was conducted by Censuswide between 25 March and 2 April 2024, with strict adherence to the Market Research Society’s code of conduct. The insights from this study come as over 60 countries worldwide are also gearing up for elections in 2024, raising universal concerns about the potential misuse of AI and deepfakes in tackling democracy.
