The Threat of AI Deepfake Robocalls in US Elections

As generative AI technology rapidly advances, so does its capacity for deception. While visual deepfakes have often been highlighted for their misleading potential, it’s audio deepfakes that pose an even greater threat in the context of elections. From Arizona to Vermont, election officials are facing the daunting task of countering fake audio calls designed to mislead voters and sow confusion. Unlike manipulated images or videos, deepfake audio is harder to detect and leaves fewer traces, making it a particularly insidious challenge.

Why Audio Deepfakes Are the New Frontier in Election Interference?

Audio deepfakes are a potent tool for misinformation because they can convincingly replicate voices, including those of political figures. This capability means that robocalls purporting to be from candidates or election officials could be used to spread false information. The subtlety of audio manipulation makes it difficult for the average voter to recognize when they are being deceived, as the human ear isn’t naturally trained to discern between authentic and AI-generated voices.

An example of this occurred earlier this year when voters in New Hampshire received robocalls impersonating President Joe Biden. These calls urged Democrats to refrain from voting in the primary, misleadingly advising them to “save your vote for the November election.” The incident, which resulted in a significant fine for the political consultant behind the scheme, was a wake-up call for election officials nationwide.

How Election Officials Are Responding?

In response to these new threats, state officials are doubling down on both modern and traditional strategies. In Colorado, election officials have been trained on what to do if they receive a suspicious call, even if it sounds like a familiar voice giving unusual instructions. The directive is simple but effective: hang up and verify by calling the known office number. This straightforward measure ensures that any orders or changes that seem out of character are immediately questioned.

Amy Cohen, the executive director of the National Association of State Election Directors, emphasizes the importance of vigilance, noting that even before the rise of AI, robocalls were a significant concern. The key now is that these calls are more deceptive than ever, increasing the workload for officials who must rely on the public’s awareness and quick reporting of suspicious activity.

Low-Tech Solutions for a High-Tech Problem:

While cutting-edge solutions are being explored, election officials are also turning to tried-and-true methods to combat disinformation. In Maine, for example, simple measures like posting signs at town centers and fire stations ensure voters receive accurate information directly. This strategy acknowledges that not everyone consumes news or updates via the internet, making physical reminders an effective means to spread the word.

Officials in Minnesota have taken it a step further by collaborating with trusted local and religious leaders. These individuals can act as community liaisons, debunking false claims and reinforcing the truth. This grassroots approach leverages the influence of respected community figures to foster trust and promote accurate information.

The Role of Media and Public Awareness:

Public awareness campaigns are essential in the fight against deepfake audio. The Illinois State Board of Elections, for example, launched an advertising campaign over the summer aimed at educating the public about the risks of election disinformation. By airing these messages on both television and radio, officials reached a broad audience, reinforcing the need for vigilance and critical thinking among voters.

In the case of New Hampshire’s January incident, rapid response was key. State officials, including the attorney general and law enforcement, issued an immediate statement denouncing the robocall. This prompt action was critical in preventing potential voter suppression and underscored the importance of being prepared for new threats posed by AI technologies.

Preparing for November 5 and Beyond:

As the US presidential election approaches, the stakes are higher than ever. Officials know that the spread of deepfake audio and robocalls can escalate in the final days before the vote, leaving little time to counteract their effects. States are committed to mobilizing all available resources, from media partnerships to on-the-ground community engagement.

This proactive approach is necessary because while deepfake audio technology might be sophisticated, it can be countered with vigilance, awareness, and strategic communication. The challenge remains significant, but with continued efforts, officials hope to safeguard the integrity of the electoral process and protect voters from the deceptive power of AI-generated deepfakes.

Conclusion:

The advent of AI deepfake technology has introduced a new level of complexity to election security. While visual deepfakes might capture headlines, it’s the audio deepfakes—often operating under the radar—that pose the most immediate threat. US election officials are tackling this issue head-on, employing a combination of advanced strategies and time-honored practices to mitigate potential damage.

The response to incidents like the fake Biden robocall demonstrates that vigilance, transparency, and public cooperation are vital. As the election draws near, these measures will play a critical role in ensuring voters are informed, protected, and empowered to exercise their democratic rights without undue influence from deceptive AI.