Americans on opposite sides of the aisle are far from opposites, sharing views on countless policy issues. But misinformation blinds us to this reality. It sharpens political division among the masses and gives us a domineering sense of schism.

Affective polarization occurs when both sides exaggerate each other’s perspectives. It feeds off of misinformation that intentionally portrays a party or demographic in a misleading way. And it works.

California’s AB 2839 is taking a step against this problem. It prohibits the distribution of election content that is “materially deceptive,” including artificially adapted content that could harm a candidate, committee or elected official’s reputation. It also prevents voting processes from being misrepresented. Though this bill might seem like a no-brainer at first, it has sparked controversy, and the state has put it on hold.

Some say it violates the First Amendment: that in trying to curb misinformation that inhibits productive conversations, we run the risk of undercutting the free speech on which those conversations depend. The problem and the solution are heavily intertwined and could become synonyms if we’re not careful.

I don’t doubt that the bill is a good first step, especially in light of AI’s increasing prevalence. At the very least, it gives our conversations something tangible to grapple with. Still, it seems to be a failed attempt at controlling misinformation. We have to continue looking for a solution that balances free speech and depolarization, the open exchange of ideas with norms that halt the spread of misinformation.

So, What’s the Bill Say, Exactly?

AB 2839 is part of a triad of bills in California that seek to combat the explosion of AI content in view of the election.

• AB 2655: requires online platforms with over 1 million users in California to identify and remove “materially deceptive content.”
• AB 2355: requires labeling AI political ads with disclaimers to inform viewers that the content was a production of AI.
• AB 2839: the only triad bill that includes civil remedies and injunctions against content creators and platforms who intentionally violate the rule, exempting political satire. It extends the enforcement period, which was laid out in AB 2655. Originally, content would be considered in violation if distributed or posted within 120 days of an election. However, this part of the bill has extended the window to include up to 60 days after. 

Why Was the Bill Halted?

Governor Gavin Newsom signed the bill into law in September 2024, making it one of the strictest laws of its kind in the U.S. However, only a month later, U.S. District Judge John A. Mendez said that, while AI and deepfakes are an issue, the bill violates citizens’ right to free speech. The bill has been on hold ever since. Mendez writes, “Most of AB 2839 acts as a hammer instead of a scalpel, serving as a blunt tool that hinders humorous expression and unconstitutionally stifles the free and unfettered exchange of ideas, which is so vital to American democratic debate.”

Some Avenues

An alternative to the bill is to increase public awareness on AI content and provide more education, helping people identify and defend against misinformation. But of course, this is not a perfect solution, if ever such a thing has existed in public policy. As AI becomes more developed and harder to spot, educational measures with limited scope and participation may not be enough. The current bill is excessive, but maybe we could use a clear definition of “materially deceptive” content and regulate content that falls under that definition. It’s possible that the bill’s lack of specificity is the problem. Using Mendez’s words, maybe we need a “scalpel,” clearly defining the limits of free speech in the age of information and AI. Unfortunately, that requires more than an op-ed.

The Future

AI content will only continue to grow, and with it, more ways to manipulate voters and stoke unnecessary, emotional division. “Me versus them” thinking and overall toxic partisanship has reached intolerable levels, driven more by misinformation, stereotyping and political messaging than substantive policy disagreement. It’s an exceedingly harmful illusion.

California has the right idea, but there is a huge risk in developing a policy that will constrict First Amendment rights. It’s important to protect Americans’ rights while ensuring election integrity, and neither should come at a cost to the other.

Acknowledgement: The ideas and opinions expressed in this article are those of the author.