AI Deepfakes Threaten Election Integrity, Raising Concerns About US Institutional Credibility
With the US election season in full swing and artificial intelligence (AI) rapidly evolving, AI manipulation in political advertising is emerging as a significant threat to election integrity and the credibility of American institutions. A new report from Moody’s on Wednesday sounds the alarm, warning that generative AI and deepfakes are among the most concerning election integrity issues.
Key Takeaways:
- Deepfakes pose a serious threat: Moody’s warns that deepfakes could be used to mislead voters, exacerbate division, and sow discord, potentially influencing the outcome of elections and undermining the credibility of US institutions.
- Government stepping up efforts: The Federal Communications Commission (FCC) has proposed a rule requiring political TV, video, and radio ads to disclose the use of AI-generated content. The Federal Elections Commission (FEC) is also considering widespread AI disclosure rules that would extend to all platforms.
- Social media companies taking action: Meta requires AI disclosure for all its advertising and is banning new political ads in the week leading up to the election. Google also requires disclosures for political ads with modified content that depicts real or realistic-looking people or events.
- Concerns about the spread of misinformation: Brands are worried about being associated with the spread of misinformation, especially as Google and Facebook are expected to capture a significant portion of digital advertising spending in 2024.
- Decentralized US election system provides some protection: Moody’s highlights the decentralized nature of the US election system, along with existing cybersecurity policies, as potential safeguards against AI manipulation.
- Public perception is crucial: Even without concrete examples of deepfakes influencing elections, the perception that they can do so can undermine public confidence in the electoral process and government institutions.
The Rise of AI-Generated Deepfakes in Political Advertising
The growing accessibility and sophistication of Generative AI tools are making it easier than ever to create convincing deepfakes. Previously, creating a deepfake required specialized knowledge and significant resources. Now, anyone can generate a sophisticated deepfake in minutes. This accessibility, combined with the volume of content on social media platforms, creates an ideal environment for the misuse of deepfakes.
Moody’s cites examples:
- Deepfake audio has already been weaponized in a presidential primary race in New Hampshire, using robocalls to mimic a candidate’s voice.
- Governments and non-governmental entities have weaponized deepfakes for propaganda, to create social unrest, and even to incite terrorism.
The Government’s Response: Regulation and Disclosure
The US government is taking steps to address the potential for AI manipulation in political advertising. The FCC’s proposed rule requiring AI disclosure in political ads is considered groundbreaking. This would be a major step toward regulating AI content in traditional media, potentially setting a precedent for future regulations in other areas of advertising.
However, the FCC’s proposal does not cover social media. The FEC is examining its own AI disclosure rules, but critics argue that it should delay its decision until after the election. They warn that requiring disclosures for traditional media while not requiring them for digital ads could confuse voters and inadvertently legitimize unlabeled AI content online.
Social Media’s Self-Policing Efforts
Social media platforms have begun to implement their own AI disclosure policies in anticipation of regulation.
- Meta requires AI disclosure for all its advertising and is banning new political ads in the week leading up to the November election.
- Google requires disclosures for political ads that use modified content to depict real people or events but doesn’t require AI disclosures for all political ads.
These efforts are seen as a step in the right direction, but they face significant challenges. The sheer volume of content posted on social media makes it difficult to catch all instances of AI-manipulated content, especially AI-generated spam messaging and AI imagery.
The Impact on Public Trust and the Electoral Process
Moody’s believes that the perception of AI manipulation, even without concrete evidence, can significantly threaten the integrity of the election process. Public distrust in the electoral process and government institutions can lead to disengagement and increased political and social risks, ultimately affecting government effectiveness.
While the US government and social media platforms are taking steps to combat AI manipulation, the rapid evolution of AI technology makes it a constantly evolving challenge. This highlights the importance of continued public education and awareness regarding the potential for AI manipulation.
Looking Forward: A Call for Vigilance and Education
As AI technology continues to evolve, the threat of AI manipulation in politics will only grow. Voters and policymakers must remain vigilant and continue to develop strategies to combat this evolving threat.
While it’s reassuring to see government and social media platforms taking action, the responsibility also falls on individuals to be discerning consumers of information. The public needs to be empowered to critically evaluate the information they encounter online and be aware of the potential for AI manipulation.
The upcoming election serves as a stark reminder of the importance of protecting election integrity and preventing AI manipulation from undermining democratic processes. The future of democracy may depend on our ability to effectively navigate these challenges.
