Big Tech, U.S. cyber officials brace for AI threats as fraught 2024 elections loom

A version of this story appeared in the Threat Status newsletter from The Washington Times. Click here to receive Threat Status delivered directly to your inbox each Wednesday.

The high-tech industry, national security officials and lawmakers on Capitol Hill are bracing for the potential for artificial intelligence applications to upend the 2024 elections, presenting unprecedented challenges for voters, candidates and election officials nationwide.

The advent of commercially available AI products brought a boon for business and entertainment last year, but its most significant impact this year may reside in what effect it has on the already fraught November elections.

One of the biggest players, search giant Google, is building its artificial intelligence tools with election integrity in mind, said company Vice President Susan Jasper.

Ms. Jasper said Google is experimenting with features for its Search Generative Experience (SGE) and Bard offerings to create faster and more adaptable enforcement systems powered by AI. Bard is Google’s conversational AI tool, or chatbot, and SGE refers to the new answers made possible via Google Search through its advanced algorithms.

“We’ve prioritized testing for safety risks ranging from cybersecurity vulnerabilities to misinformation and fairness,” Ms. Jasper wrote on the company’s blog in December. “Beginning early next year, in preparation for the 2024 elections and out of an abundance of caution on such an important topic, we’ll restrict the types of election-related queries for which Bard and SGE will return responses.”

Precisely what questions Google chooses to censor about the elections remains to be determined. Google-owned YouTube will require creators to disclose whether they have used AI to make or alter content and provide a warning label for their audience.

Ms. Jasper said Google is working with the nonprofit Democracy Works to decide what information to put atop search results for people seeking information from state and local election offices on when, where and how to vote. Early results from the company’s experiments suggest that Google’s response time to emergent threats will improve as a result of AI, Ms. Jasper said.

The company’s cybersecurity specialists already have their hands full monitoring more than 270 cyberattack groups across more than 50 countries, some of which have the support of foreign governments eager to sow divisions among the American electorate. Russia, China and Iran are some of the adversaries that U.S. officials say have actively tried to skew the U.S. political debate online.

The number of hackers and online users conducting influence and information operations to manipulate American voters is growing all the time, officials say. The reduced cost of advanced computing tools and knowledge of how to deploy them have lowered the barrier to entry for hostile hackers around the globe.

The problem isn’t just for the United States. Nations around the world, including the populous Brazil, Russia and India, will hold elections this year. Security analysts say Taiwan’s national elections on Jan. 13, in the face of an increasingly aggressive pressure campaign from China, will pose an early test of AI and its threat to democratic electoral systems.

Sam Altman, CEO of the market-leading firm OpenAI, issued a stark warning about the challenges during a panel discussion in November on the sidelines of the Asia-Pacific Economic Cooperation summit in San Francisco.

“The dangerous thing there is not what we already understand … but it’s all the new stuff — the known unknowns, the unknown unknowns,” he said. “There’s a whole bunch of other things that we don’t know because we haven’t all seen what, you know, generative video or whatever can do, and that’s going to come fast and furious during an election year.”

A road map from 2022

The threats encountered during the 2022 midterm elections provided a road map for America’s digital defenders to follow and uncover further incursions.

Chinese government leaders aimed at Congress as state-linked hackers scanned more than 100 state and national party web domains, according to an intelligence community assessment published in December.

Iranian hackers masqueraded online as left-leaning American voters, expressed support for liberal politicians, and made plans to deploy covert social media operations and fraudulent news agencies to manipulate American reporters.

Russia’s active measures were carefully crafted and aimed at specific audiences such as “U.S. males who were more than 40 years old and interested in ‘right-wing conservatism,’” according to the report.

All of the activity two years ago has America’s cyberwarriors on guard. Army Gen. Paul Nakasone, outgoing leader of NSA and Cyber Command, mobilized the election security group at the National Security Agency and U.S. Cyber Command to combat foreign threats.

Gen. Nakasone said last month that the agencies under his watch are fixated on answering questions about whether China will change its disruption playbook and how enhanced AI capabilities will worsen the threat this year.

Air Force Lt. Gen. Timothy D. Haugh, confirmed by the Senate last month as the next head of the NSA and Cyber Command, worked on election defense efforts in each U.S. electoral cycle dating back to 2018. In July, Gen. Haugh told lawmakers he was anxious about the foreign use of AI tools to disrupt the electoral process.

Microsoft threat specialists wrote in a November analysis that “presidential elections determine the course of foreign policy and for authoritarian nation-states — principally Russia, Iran, and China — next year’s presidential contest will be critical for each of these countries seeking to advance their strategic goals.”

The U.S. vote “may be the first presidential election during which multiple authoritarian actors simultaneously attempt to interfere with and influence an election outcome …,” the Microsoft study said. “Election defense efforts — from both the U.S. government and the private sector — will likely be heavily tested this year and next as the election cycle begins.”

Lawmakers say they hear alarm from the technology industry and intelligence community. Senators have identified AI risk to elections as a leading priority for any forthcoming cybersecurity legislation, said Sen. Todd Young, Indiana Republican.

Mr. Young has worked hand in hand with Senate Majority Leader Charles E. Schumer, New York Democrat, to draft legislation focused on AI. In October, Mr. Young said AI’s impact on votes must be addressed regardless of how the elections play out.

With former President Donald Trump the clear front-runner for the Republican nomination, party voters’ appetite for accepting Democratic-led AI legislation is small and may shrink as the elections approach. R Street Institute senior fellow Adam Thierer said voters will be understandably suspicious of their political opponents’ motivations to govern AI’s effects on elections.

“Most conservatives believe a lot of algorithms and tech companies are already stacking the decks against them in other ways,” Mr. Thierer said.

“Well, now Joe Biden and Chuck Schumer are going to come in here and tell us more about how this is supposed to work. That debate’s going to be neutralized because of that fighting, that fight over that core issue. I don’t think anything happens there at the end of the day.”