31 States Regulate Election AI, 19 Don’t: What’s Next Thirty-one U.S. states had enacted laws regulating AI in election campaigns as of October 2026 while 19 states had none, according to SciLine figures cited in a Law.com-framed analysis, and no federal regulation specifically governs AI in political advertising. At the federal level, the Federal Election Commission declined to write a dedicated AI rule and the Federal Communications Commission banned AI-generated voices in robocalls in 2024, leaving state deepfake and labeling requirements plus older statutes as the only guardrails. The analysis found no confirmed federal AI law, regulation or executive action tied to the election, so compliance teams and campaign vendors should plan around the 31-state patchwork rather than an unconfirmed federal overhaul. Thirty-one states have laws on AI in election campaigns, 19 have none, and the United States still has no federal rule written for AI political advertising. SciLine reported those figures as of October 2026, and they frame the debate Law.com headlined this week as Rewiring Policy: What the Election Means for Artificial Intelligence and Technology. This analysis separates what the public record confirms from what the headline only implies. The AI industry has watched this cycle closely. OpenAI chief Sam Altman told Bloomberg House in Davos in January 2024, in remarks reported by CNBC https://www.cnbc.com/2024/01/19/openais-sam-altman-ai-will-be-fine-regardless-of-us-election-outcome-.html , “I believe that America is gonna be fine, no matter what happens in this election. I believe that AI is going to be fine, no matter what happens in this election, and we will have to work very hard to make it so.” He said that about the 2024 race, so treat it as history and not as a comment on today’s vote. Pressure on AI companies also comes from courts, not only ballots. Our look at whether an unfair competition suit can freeze frontier AI https://shattered.io/frontier-ai-lawsuit-unfair-competition-law-injunction-five-labs-zero-rulings-2026/ shows how quickly the legal front is widening. Below, we cover the facts, the gaps, the compliance math and five predictions for what comes next. Don't miss new tech stories on Google Add Shattered once in the Google app and our stories show up in your news suggestions. Add now https://www.google.com/preferences/source?hl=en&q=https%3A%2F%2Fshattered.io%2F The Core Facts: 31 States With Rules, 19 Without Start with the numbers that hold up. According to SciLine, 31 states had enacted laws regulating AI in election campaigns as of October 2026. Those laws mostly target deepfakes and require labels on AI-generated content. The same SciLine reporting says 19 states had no such laws on the books. At the federal level, SciLine reports that the country had no regulations specifically governing AI in political advertising. That does not mean federal law is silent. Two agencies acted in 2024, and both actions still shape how campaigns, broadcasters and AI vendors behave. The Federal Election Commission declined to write a dedicated AI rule. The Federal Communications Commission banned AI-generated voices in robocalls. Put plainly, the rulebook is a patchwork. A campaign in one state faces labeling duties and possible deepfake limits. A campaign in a neighboring state may face neither. The only nationwide guardrails are old statutes stretched to cover new tools. That is the baseline any new Congress, administration or agency starts from, whatever the election’s outcome turns out to be. Why the Election Puts AI Policy Back in Focus Elections concentrate attention. Every cycle produces a fresh batch of synthetic audio, doctored video and AI-written messaging, and each incident becomes an argument for or against new rules. AI election policy rarely moves in a straight line. It lurches after a viral fake, then stalls while lawyers argue about speech rights. The Law.com framing suggests the vote may reshape that dynamic. We could not match the headline to a verified underlying article, and our sources do not establish which election it covers or what result it produced. We also found no confirmed federal AI law, regulation or executive action tied to the vote. So the honest read is narrow. The election raises the political value of AI rules, but the record does not yet show a specific policy change. That distinction matters for readers who make decisions on this story. Compliance teams, campaign vendors and AI product managers should plan around the 31-state patchwork and the 2024 agency positions. They should not plan around a federal overhaul that no source has confirmed. What Federal Agencies Have Actually Done Two agencies hold the federal levers that touch AI in campaigns today. Their approaches differ sharply, and the difference explains most of the current gap. The FEC’s Technology-Neutral Position In 2024, the Federal Election Commission declined to create a proposed federal rule on deceptive AI campaign advertisements. It concluded that its existing prohibition on fraudulent misrepresentation is “technology neutral.” In practice, the agency says a lie in a campaign ad is a lie whether a human editor or a generative model produced it. The FEC confirmed that its fraudulent-misrepresentation provisions apply to AI-generated communications. It did not issue new AI-specific rules, and it did not require extra AI disclaimers. That choice favors case-by-case enforcement over a bright-line statute. Questions about definitions, proof and penalties stay open until a complaint forces an answer. The FCC’s Robocall Ban The Federal Communications Commission took a narrower but firmer step in 2024. It banned AI-generated voices in robocalls, including campaign-related robocalls, under the Telephone Consumer Protection Act. The rule works because it targets a delivery channel with an existing legal hook. Voice cloning over the phone fell inside a statute written decades earlier. The contrast is useful. The FCC acted on a channel it already regulates. The FEC declined to act on a category, AI ads, that its law does not neatly address. Neither agency has a mandate to police AI-generated video on social platforms, which is where much of the deceptive content spreads. How State Deepfake and Labeling Laws Work States filled the vacuum. SciLine’s count of 31 states shows that this is no longer a handful of early movers. Most of these laws aim at two things: deepfakes of candidates and labels on AI-generated content. They differ on timing windows, penalties, and who can sue. State laws largely target synthetic media and not general-purpose AI. A law that requires a disclosure on a manipulated video says nothing about the model that produced it. That design keeps the statutes narrow enough to survive constitutional scrutiny, and it leaves chatbots, recommendation systems and AI-written text mostly untouched. The 19 states without laws are not necessarily indifferent. Some may rely on older fraud or defamation statutes. Others may have bills pending. Our sources do not say which, so the safest statement is the one SciLine made: those states had no laws regulating AI use in elections as of October 2026. The patchwork has a second effect on AI developers. When rules vary by state, model providers cannot ship a single election-safety policy. They tend to adopt the strictest common denominator, such as watermarking or refusal behavior on candidate likenesses, because it is cheaper than tracking 31 rulebooks. There is also a timing problem. Election-window rules only help if they apply before voters see the content. A label added after a fake clip has gone viral does little to undo the damage, and a takedown that lands a week late may arrive after ballots are cast. Lawmakers who draft these statutes know this, which is why many focus on windows before an election and on fast civil remedies for targeted candidates. Enforcement is the quieter weakness. A law on the books does not guarantee a prosecutor, an election office or a judge will act quickly. Our sources give no enforcement counts for the 31 states, so we cannot say how often these laws have been used. Readers should treat the count of enacted laws as a measure of legislative intent, not of real-world deterrence. Regulators Compared: Scope, Tools and Gaps The table below lines up the main players using only the confirmed record. It shows where each one has authority and where the open gaps sit. | Actor | Action on record | Year | Main gap | |---|---|---|---| | Federal statute on AI political ads | None specific to AI in political advertising | As of Oct. 2026 | No national definition or disclosure rule | | Federal Election Commission | Declined a dedicated AI rule, called its fraud provision technology neutral | 2024 | No AI-specific disclaimers required | | Federal Communications Commission | Banned AI-generated voices in robocalls under the TCPA | 2024 | Covers calls, not social video | | States with election AI laws | 31 states regulate deepfakes and AI labeling | As of Oct. 2026 | Rules differ state to state | | States without election AI laws | 19 states have no such law | As of Oct. 2026 | Reliance on older statutes | | Existing campaign fraud rules | Apply to AI-generated communications per the FEC | 2024 | Enforcement is case by case | Read the table by row, and the pattern is plain. Federal action is thin and channel-specific. State action is broad in count but inconsistent in content. Nobody owns the platform layer, where synthetic video and images circulate fastest. That gap is why the preemption question keeps surfacing. If a single federal standard arrives, it would replace the left-hand column’s inconsistency with one rule. If it does not, the 31-to-19 split probably persists through the next cycle. What Is Confirmed and What Is Not News about elections and technology attracts overstatement, so we draw a hard line between verified and unverified claims. The second table lists the main items. | Claim | Status | Basis | |---|---|---| | 31 states regulate AI in election campaigns | Confirmed | SciLine, as of Oct. 2026 | | 19 states have no such laws | Confirmed | SciLine, as of Oct. 2026 | | No federal rule specific to AI political ads | Confirmed | SciLine, as of Oct. 2026 | | FEC declined a dedicated AI rule in 2024 | Confirmed | SciLine | | FCC banned AI voices in robocalls in 2024 | Confirmed | SciLine | | Which election the headline covers | Unconfirmed | No matching verified source | | The election produced a federal AI law, rule or order | Unconfirmed | No source found | | Specific vote totals, officials or quotes from the underlying story | Unconfirmed | No verified article or on-record quote | Anything in the bottom three rows should carry a caution label until a primary source appears. We will update this analysis if Law.com or another named outlet publishes details that settle them. The Compliance Problem for Campaigns, Platforms and Broadcasters A campaign that buys ads in several states now needs a map of labeling rules. A broadcaster that airs those ads must decide whether it is on the hook for a missing label. A platform that hosts them must decide how fast to remove a flagged deepfake. Each actor sees the patchwork from a different angle, and each pays for it differently. For campaigns, the cost is mostly legal review and creative workflow. Teams that use AI to generate images, voice or copy have to track which assets need disclosure in which states. A mislabeled ad risks complaints, takedowns and news cycles the campaign did not want. For platforms, the cost is moderation at scale. The FEC has not told platforms to add AI disclaimers, and the FCC’s ban covers calls. So platform policies end up driven by state law and by reputational risk. Expect internal rules that go further than any statute demands. For AI vendors, the exposure is reputational more than legal. A model that clones a candidate’s voice on request invites headlines and regulator attention, even when the law does not clearly reach it. That pushes vendors toward refusal policies on named political figures during election windows. Federal Preemption: The Fight Ahead Preemption is the issue most likely to define the next phase. A federal framework could set uniform definitions and disclosure rules. It could also displace stricter state requirements, which is why states and advocacy groups resist it. The election raises the stakes of that conflict because both sides can point to fresh examples. The argument also overlaps with the trade side of AI policy. Washington has recently tied technology rules to supply chains and alliances, as in the Pax Silica pact signed by the US and eight allies https://shattered.io/us-eight-allies-sign-pax-silica-san-francisco-declaration-ai-supply-chains/ . Our explainer on the China AI export controls draft https://shattered.io/what-the-china-ai-export-controls-draft-actually-says/ shows the same instinct at work. Federal power tends to flow toward national-security framing first and consumer or election framing second. Election law adds a wrinkle. Under the constitutional structure, states run federal elections, and Congress keeps authority to regulate aspects of them. That split means a preemption bill on AI political ads would need to be written carefully to survive challenge. Speech concerns add a second layer, since campaign messaging sits near the core of protected expression. A practical point for readers who build or buy AI tools: preemption fights rarely resolve quickly. Even if a bill moves, definitions of terms like synthetic media, materially deceptive and political communication would take months of drafting and comment. Companies that wait for clarity before acting will spend the interim exposed to the existing patchwork, so the cheaper path is to build disclosure and provenance features now. The politics cut across party lines in unusual ways. Some lawmakers favor strict deepfake limits because they fear being targeted. Others favor light rules because they fear overreach on speech. A federal bill would need to satisfy both camps, and that tension helps explain why the FEC chose not to write a rule in 2024 and why Congress has not filled the space since. Historical Context: From 2023 Testimony to 2024 Rules The current patchwork did not appear overnight. In May 2023, Altman told a Senate panel that he was worried about AI-driven election interference. Reuters covered the hearing https://www.reuters.com/technology/openai-chief-goes-before-us-congress-propose-licenses-building-ai-2023-05-16/ , and Altman’s short answer on the topic was “I am nervous about it.” The remark came a year before the agencies below made their calls. By 2024, federal agencies had to decide how far existing law stretched. The FCC moved first on a channel it already controlled. The FEC took the opposite route and declined to write a new rule. Together, those decisions told Congress that agencies would not build an AI election framework on their own. The state count grew in the gap. SciLine’s tally of 31 states shows how far the legislative wave traveled. The tally also shows its limit, since 19 states still had no law as of October 2026. In short, two years after the federal agencies made their calls, the country is still waiting for a national answer. What AI Labs Say About Election Policy, and Why Dates Matter Industry voices have mostly argued for balance. Chris Lehane, OpenAI’s Chief Global Affairs Officer, told Semafor in October 2024 that the central task is how to strike and get to the right balance between deploying AI safely and keeping the country positioned to lead. His exact words, in Semafor’s interview https://www.semafor.com/article/10/30/2024/sam-altmans-political-guru-on-the-us-election-and-democratic-ai-vs-autocratic-ai , were: “It really is how you strike and get to the right balance between ensuring that AI is deployed in a safe and responsible way, while also making sure that the country is poised to leverage its leadership role.” All three quotes in this article come from 2023 and 2024. They describe how AI executives framed the issue before this cycle, and none of them reacts to the current election. We cite them because they show a consistent industry line: accept some rules, resist rules that slow US competitiveness, and avoid being the face of an election scandal. That line is under strain. Safety groups and some lawmakers argue that voluntary company policies do not substitute for law. Our coverage of Anthropic’s decision to say that its responsible scaling policy threshold was met https://shattered.io/anthropic-responsible-scaling-policy-threshold-met-ipo-1-4-trillion-2026/ shows that even labs now talk openly about risk tiers. Whether election misuse sits inside those tiers is a question regulators have not answered. Market Impact: Who Pays for the Patchwork We have no confirmed dollar figure for the cost of election AI compliance, so we will not invent one. The structure of the cost is clear, though. Fragmented rules favor large companies with legal teams and penalize small vendors that sell AI creative tools to campaigns. Winners. Compliance software firms, large platforms with trust-and-safety staff, and AI vendors with strong content provenance tools stand to gain. A patchwork creates demand for tools that tag, trace and label synthetic media automatically. Losers. Small campaign vendors and local consultants carry the heaviest relative burden. They lack in-house counsel, and a single mislabeled asset can cost them a client. Broadcasters in states with strict rules face similar uncertainty about their own liability. Investors should also watch sentiment. The AI trade has been sensitive to regulatory headlines, and an election that revives federal preemption talk or new disclosure mandates can move expectations for model providers. Our report on Nvidia’s slide before earnings https://shattered.io/nvidia-down-30-percent-ai-selloff-resumes-before-earnings/ shows how quickly policy and market mood can merge. Five Predictions for AI Policy After the Election These are our judgments, not confirmed events. Each rests on the confirmed record above and could change if new facts emerge. 1. The 31-state count rises. With 19 states still uncovered, sponsors have an obvious target list for the next legislative session. 2. Federal preemption language appears in at least one bill. The compliance burden gives industry a strong argument, and the election gives lawmakers a reason to move. 3. The FEC keeps its technology-neutral stance. Absent a new commission majority or a court order, we expect case-by-case enforcement to continue. 4. Platforms tighten their own election AI policies. Voluntary rules fill the gap fastest, especially on candidate likenesses and voice cloning. 5. Courts shape the boundary. Challenges to state deepfake laws will test where political expression ends and regulable deception begins. None of these predictions assumes a specific election result. They hold across most plausible outcomes because they follow from the structure of the rules, not from who won. What to Watch Next Four signals will tell readers whether the headline’s promise of rewired policy turns real. First, any new federal bill that mentions AI in political advertising. Second, any FEC or FCC action that changes the 2024 positions. Third, new state enactments that move the 31 count. Fourth, court rulings on deepfake statutes. Outside elections, the pressure on frontier AI keeps building. Nonprofits have taken the fight to court, as we covered in our report on the suit against five AI developers https://shattered.io/nonprofits-sue-five-ai-developers-halt-frontier-ai-63-page-complaint-2026/ , and Europe has issued its own call in the Dublin AI Declaration https://shattered.io/dublin-ai-declaration-seven-priorities-epc-leaders-europe-2026/ . For more background on models and policy, see our AI and machine learning section https://shattered.io/ai-machine-learning/ . The takeaway: the verified record shows a country with 31 state rulebooks, 19 gaps and two narrow federal actions from 2024. A new federal framework is possible, but no source we found confirms one. Readers should plan for the patchwork and watch for the signals above. FAQ: AI, Technology and the Election Does the US have a federal law on AI in political ads? No. According to SciLine, the United States had no federal regulations specifically governing AI in political advertising as of October 2026. How many states regulate AI in elections? SciLine reports that 31 states had enacted laws regulating AI in election campaigns, mainly on deepfakes and labeling. Another 19 states had no such laws. What did the FEC decide about AI campaign ads? In 2024, the Federal Election Commission declined to create a dedicated rule on deceptive AI campaign advertisements. It said its existing fraudulent-misrepresentation prohibition is technology neutral and applies to AI-generated communications. It did not require new AI disclaimers. What did the FCC ban? In 2024, the Federal Communications Commission banned AI-generated voices in robocalls, including campaign-related ones, under the Telephone Consumer Protection Act. Did the election produce a new AI law or executive order? We found no confirmed evidence of one. Our sources do not establish a specific federal AI law, rule or executive action tied to the vote. Which election does the Law.com headline cover? Our sources do not confirm that. We could not match the headline to a verified underlying article, so this analysis avoids claims about results or vote totals. Why does federal preemption matter? A single national standard could replace inconsistent state rules and cut compliance costs. It could also override stricter state protections, which is why the issue draws opposition.