Who’s behind the Catherine Connolly deepfake video, and why?

The Incident: Ireland’s First Electoral Deepfake

Earlier this week, Irish social media users encountered a strikingly realistic video that appeared to be from RTÉ News, one of the country’s most trusted broadcasters. The video falsely announced that the election had been cancelled and that Heather Humphreys had been declared the winner because Catherine Connolly had supposedly withdrawn from the race. In reality, none of these claims were true.

This video was shared widely on platforms like Facebook and YouTube, marking Ireland’s first known use of a sophisticated deepfake in an electoral context. The content’s convincing nature raised alarms about the potential impact on voter behavior and public trust in democratic processes.

Possible Motivations Behind the Deepfake

To understand why someone would create and distribute such a deepfake, it is important to consider the various motivations that typically drive electoral disinformation campaigns.

One primary reason is to influence the outcome of an election by misleading voters. Disinformation about voting procedures, eligibility, or election status can suppress turnout among specific groups. For example, in the United States during the January 2024 Democratic primary in New Hampshire, AI-generated robocalls impersonating Joe Biden falsely instructed voters to skip the primary, aiming to reduce support for Biden and advantage his opponent.

However, in the Irish case, the deepfake video was not evidently targeted at any particular demographic or voter group. This suggests that directly influencing the election outcome may not have been the primary goal.

A second, and perhaps more plausible, motivation is to undermine democracy itself by sowing confusion and distrust. Historical examples include Russian interference in the 2016 US Presidential election, where the primary objective was to create discord and weaken public faith in the electoral process rather than to support a specific candidate initially. Similarly, Romania’s 2019 presidential election was disrupted by online interference that led to the election’s cancellation and rerun, damaging trust in democratic institutions.

A third possibility is that the deepfake was created as a marketing demonstration by a deepfake technology provider. There is a growing industry of firms offering sophisticated disinformation tools to political campaigns and other clients, both legitimate and illicit. High-profile firms like Cambridge Analytica previously used controversial online tactics to influence elections and leveraged media attention to showcase their capabilities. Smaller, less visible operators also benefit from demonstrating their technology through high-profile examples to attract customers. The Irish deepfake video could be a deliberate showcase intended to advertise the capabilities of a deepfake vendor.

Finally, the deepfake might have been produced as a form of nihilistic trolling or an online 'flex'—a way for creators to display their technical skills or provoke reactions without a clear political agenda. The internet has a history of such pranks, including the promotion of fabricated events simply for amusement or to test public gullibility.

Broader Implications for Democracy and Public Trust

The emergence of deepfake technology in electoral contexts poses serious challenges to democratic societies. The ability to create highly convincing fake content that appears to come from trusted sources can erode confidence in legitimate news and official communications.

When voters cannot easily distinguish real information from fabricated content, their trust in the electoral process and its outcomes may diminish. This can lead to voter apathy, increased polarization, and a weakening of democratic norms.

Moreover, the possibility that such deepfakes are used as marketing tools by disinformation vendors complicates efforts to combat electoral interference. It suggests a commercial incentive to develop and distribute increasingly sophisticated fake content, potentially escalating the arms race between disinformation creators and defenders.

The case in Ireland, while not on the scale of interference seen in larger elections, serves as an early warning. It highlights the need for proactive strategies to detect, expose, and counter deepfake disinformation before it can cause significant harm.

Responding to Deepfake Disinformation: What Can Be Done?

Addressing the threat of electoral deepfakes requires a multi-faceted approach involving technology, policy, education, and public awareness.

Technological solutions include the development of advanced detection tools that can identify manipulated videos quickly and reliably. Social media platforms must also enhance their monitoring and moderation systems to limit the spread of harmful deepfakes.

Policy measures could involve clearer regulations on the use and dissemination of AI-generated content, especially during sensitive periods like elections. Transparency requirements for political advertising and content sponsorship may help reduce the impact of covert disinformation campaigns.

Public education is equally important. Voters should be informed about the existence and risks of deepfakes and equipped with critical media literacy skills to evaluate the authenticity of information.

Finally, fostering collaboration between governments, technology companies, civil society, and independent media can create a more resilient information ecosystem capable of resisting manipulation.

What this means

The appearance of a deepfake video in Ireland’s electoral context signals a new frontier in the challenges democracies face from AI-driven disinformation. While the precise motives behind this particular incident remain unclear, the implications are undeniable: sophisticated fake content can disrupt trust and potentially influence political outcomes. Recognizing the multifaceted motivations behind such acts is essential to crafting effective responses. As deepfake technology continues to evolve, so too must the vigilance and resilience of democratic societies. Preparing for future incidents involves not only technical innovation but also informed public discourse and robust policy frameworks to safeguard the integrity of elections and public trust.