Deepfake research warns Scottish Parliament TV vulnerable to attacks
Research from the University of Edinburgh and Scottish Centre for Crime and Justice Research identified deepfake threats to Scottish Parliament TV, which livestreams and archives Holyrood proceedings publicly. A May 2024 deepfake video of First Minister John Swinney circulated online, and the research warned that publicly accessible parliamentary videos could be manipulated to spread false statements by MSPs.
What happened
Research commissioned by the Scottish Parliament from the University of Edinburgh and Scottish Centre for Crime and Justice Research identified vulnerabilities in Parliament TV to deepfake attacks. The researchers warned that deepfake technologies could be used to create convincing fake videos of MSPs saying things they never said, potentially spreading through social media.
In May 2024, a deepfake video of First Minister John Swinney appeared on social media, putting false words in his mouth. Former First Minister Humza Yousaf, Green party co-leader Patrick Harvie and Green MSP Maggie Chapman were also targeted with doctored videos.
Reported impact
The research identified three key threats to Parliament TV: hackers could intercept the live stream by breaking into network encoders and inserting manipulated video, attackers could target people dialling in via Zoom with live deepfakes, and the broad accessibility of archived videos on Parliament TV makes them highly susceptible to manipulation and reuse.
Dr Ben Collier of SCCJR and the University of Edinburgh warned that while livestreaming was originally intended to produce a neutral public record of parliamentary proceedings, it has paradoxically introduced new security risks that could threaten public trust in democracy.
Index analysis
Scottish Parliament TV is one of the world's first legislatures to introduce public livestreaming and archiving of parliamentary proceedings. The Parliament TV website has no content restrictions and allows users to download video clips using a dedicated 'clipping' tool with broadly permissive reuse licenses, making the material easily accessible for deepfake creation.
A precedent for such attacks exists: in 2024, Iranian armed forces hacked into streaming services run by the United Arab Emirates with a broadcast featuring an AI-generated news anchor, demonstrating the technical feasibility of compromising parliamentary video streams.
Governance controls
The research team found many examples of good practice and a high level of resilience overall, but called on the Scottish Parliament to put a formal process in place to address deepfake threats. The research recommended improvements to secure Parliament TV against deepfake attacks.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.