Artificial intelligence (AI) makes it easier for malicious actors to spread convincing election disinformation and launch cyberattacks, but the technology also offers new ways to strengthen election security, according to Gary Barlet, Illumio’s public sector chief technology officer.
In an interview with MeriTalk, Barlet warned that AI-generated falsehoods pose a particularly serious challenge, while vulnerabilities in the systems that collect, store, and transfer election results also warrant greater attention.
“[AI] is becoming more and more prevalent, across the board. It’s going to be a significant benefit and a significant threat to election security. This election and all elections moving forward,” Barlet said.
AI-generated disinformation presents growing risks
Barlet warned that widely available AI tools make it easier to create realistic fake audio and video that can quickly spread false information about elections.
Gary Barlet, Illumio’s Public Sector CTO
“You don’t have to be an expert to use any of these tools,” he said.
To counter the threat, Barlet urged AI developers to watermark synthetic content, social media platforms to label manipulated material, and election officials to respond quickly to false claims.
He also said federal policymakers have focused more on AI safety and development than on disinformation.
“I’m not sure they’ve really gotten to that disinformation part of the equation, and I don’t know when they’re going to get to that,” he said.
The United States has no comprehensive federal law regulating AI-generated misinformation or deepfakes, but lawmakers are considering transparency measures. Reps. Valerie Foushee, D-N.C., Don Beyer, D-Va., and James Moylan, R-Guam, introduced the bipartisan Protecting Consumers from Deceptive AI Act, which would establish technical standards for identifying and labeling AI-generated content, rather than banning deepfakes or requiring disclosures. The bill has advanced through a House committee but has not passed the full House.
While most conversations around AI and disinformation center on deceptive use of the technology, Barlet said AI could help voters verify claims by tracing information to its original source and comparing it across independent outlets. He emphasized that such tools must be easy for voters to use.
“I don’t think AI is going to be the end of the world. I don’t think AI will end democracy. I don’t think AI is going to be the end of fair elections,” he said. “Because in totality, I think AI is going to be more of a benefit than it’s going to be a detriment.”
Interconnected election systems need stronger protection
Beyond disinformation, Barlet warned that systems used to collect and transfer election results could be vulnerable to cyberattacks, particularly where data from multiple locations is consolidated.
“A lot of elections happen at the local level, where results are tabulated in Excel spreadsheets or small databases before being transferred elsewhere for final tabulation,” Barlet said. “That’s where I’m most worried, because that’s a concentration of data.”
To reduce those risks, Barlet recommended that election officials identify critical systems, understand how they connect, and limit unnecessary access, including direct internet connections.
“They should take those collection points, those critical assets, and they should have an understanding of what other systems they’re talking to, but more importantly, put containment boundaries around those systems to ensure that they’re only talking to the things they absolutely need to be talking to,” he said.
Barlet also encouraged election cybersecurity teams to use AI-powered defensive tools alongside basic security measures, noting that AI can make cybersecurity capabilities more accessible to both attackers and defenders.
“Just like AI is helping the attackers not have to be as technical, it’s also helping the defenders to not necessarily have to be as technical as well,” Barlet said.
Federal support and intelligence sharing remain critical
Barlet also called for additional federal assistance to help state and local election authorities strengthen their cybersecurity capabilities. Many local jurisdictions face staffing and budget constraints that limit their ability to implement sophisticated protections independently, he said.
“The federal government has a role to play when it comes to providing capabilities and providing funding” that would allow election officials to implement necessary safety guardrails at the local level, he said.
Barlet said federal agencies could also improve how they share intelligence about potential election-related cyber threats with state and local officials. He also emphasized that effective election security requires coordination across government, technology companies, and election stakeholders.