The Alan Turing Institute has released a study on public perception of deepfakes in the UK, revealing widespread concern about their potential impact on elections and society at large. The nationally representative survey of 1,403 UK residents found that 87.4% are worried about deepfakes affecting election outcomes, while 91.8% express concern about their broader spread.
Key findings from the survey include:
- Nearly half (49.3%) of respondents have seen non-harmful deepfakes created for educational or entertainment purposes.
- On average, 15% of people have been exposed to harmful deepfakes, including deepfake pornography, frauds, and scams.
- 50.2% have seen celebrity deepfakes online, compared to 34.1% who have seen deepfakes of politicians.
- Men are more likely to report seeing deepfakes, while women are more concerned about becoming targets of harmful deepfakes.
- Despite high awareness, most people lack confidence in their ability to detect deepfakes, with nearly 70% still trusting audio and visual content online.
Tvesha Sippy, lead author of the study, emphasised the need for a concerted effort to tackle the spread of fake content and improve media literacy. Dr. Jonathan Bright, head of online safety at the Institute, stressed the importance of source-checking and awareness of potential inauthenticity in online audio and video content.
The study also highlighted public preferences for addressing the issue, with most respondents favouring bans on users creating harmful content and improved reporting mechanisms on platforms.
As deepfake technology continues to advance, this study underscores the urgent need for regulators and society to address the potential threats to democratic processes and public trust in information. The findings call for increased efforts in media literacy and the development of effective strategies to combat the spread of harmful deepfakes.