Election Tech

Lawmakers in California have proposed legislation on everything from deepfakes to data transparency to large system safeguards. The big question: which bills could pass?

From robocalls to deepfakes, misuse of artificial intelligence threatens to play a significant role in the upcoming presidential election. Legislation from lawmakers nationwide proposes to keep it in check.

The San Francisco company described in a blog post how its chatbots have already been misused to post political spam and generate entire articles. It also described steps it is taking to stop these practices.

Signed by the state’s governor, the new law criminalizes distributing “materially deceptive media.” A first violation would be a misdemeanor, but subsequent violations would be treated as felonies. The law takes effect Oct. 1.

The IT-ISAC has started encouraging election system vendors and security researchers to collaborate on finding and fixing vulnerabilities, with a new event planned for 2025.

More than three-fourths of Americans expect that abuses of artificial intelligence will affect the 2024 presidential election, according to a new national survey.

The tool, which is called the Ballot Verifier, will give its users free and direct access to every ballot cast in Ada County, Idaho, since 2022, said the Ada County Clerk’s Office in April.

Sponsored by the Michigan Secretary of State's Office, the closed-session event took place at the Michigan Works office in Traverse City, with more sessions planned elsewhere in the coming weeks.

Secretary of State David Scanlan said he’s in favor of using federal election funds, via grants, to help cities and towns modernize aged devices. Scanlan had been reluctant to commit federal Help America Vote Act funds to the effort.

Experts point to the swift rise of artificial intelligence and generative AI, and its early use in political campaigns, as evidence it could be “weaponized to deceive voters or harm candidates” during the general election this fall.

Lawmakers are considering laws that would let officials reduce the number of voting machines and put pictures of all ballots online. Others would criminalize deepfake campaign ads and eliminate using ballot QR codes to count votes.

Elections officials have deployed new voting machines at three vote centers to better accommodate people with disabilities, and a way for voters to “cure” or fix signature problems via text message. In-person voting began Monday morning.

In the face of concerns like AI-powered phishing, tensions around discussing misinformation and physical threats, election workers can turn to several organizations aimed at providing them with help.

In the midst of an election year, evolving AI has in part led to a massive spike in deepfake-powered disinformation, but at a recent Brookings event, experts discussed how lawmakers and officials can play defense.

Committee on House Administration Chairman Bryan Steil queried Nevada Secretary of State Cisco Aguilar on a state website malfunction. The site showed mail-in ballots had been submitted for the presidential primary when they hadn’t.

With questions arising around election trust and security, some experts are proposing that U.S. voting machines shouldn’t use proprietary software, instead moving toward an open source model.

Some voters Tuesday in Tarrant County, Texas, waited in long lines to cast their ballots. The county is working with KNOWiNK, which provides its electronic poll book system, to provide wait times to the public online.

A senior official said that a widespread social media outage on Tuesday appeared to be unrelated to the elections. By the afternoon, the disruption had mostly ended.