The FCC Is One Ruling Away From Legalizing AI Robocalls to Your Cellphone, Days Before the Midterms
Eight days before the 2026 midterms, the Federal Communications Commission is sitting on a decision that could put AI-generated political pitches directly into tens of millions of cellphones without anyone's consent. The request didn't move through a multi-year rulemaking docket. It arrived as a waiver petition, a narrower and faster instrument, and the three-member commission now has what it needs to rule at any moment.
The petition comes from Club for Growth, a conservative advocacy group whose president, David McIntosh, argues that political calls to cell phones deserve the same treatment the Telephone Consumer Protection Act already gives calls about package deliveries, bank alerts, and doctor's appointments. His framing is explicitly about parity: landline political robocalls are legal without consent; wireless ones currently aren't. The petition proposes its own guardrails, capping calls at three per recipient every 30 days and promising a 10-day window to process opt-outs.
The cap that isn't a cap
Those numbers sound like a limit until you notice what they actually constrain: one caller's behavior, not a voter's total exposure. Thousands of distinct campaigns, parties, PACs, and the dark-money groups that fund independent expenditures would each get their own three-call allotment to the same phone number. An attorney with the National Consumer Law Center has put the realistic ceiling at roughly a fortyfold jump in political call volume, approaching 15 billion calls across a single election cycle. A per-caller cap that doesn't bound the number of callers is not really a cap on what a voter's phone experiences in the final stretch of a campaign.
A clock built to run out
The procedural timeline tells its own story. Comment and reply windows closed in early-to-mid October, and the commission signaled a ruling could come as early as late October, on a compressed schedule explicitly tied to affecting the November 3 election rather than following it. Whatever legal backstops survive a waiver, state deception statutes, federal anti-intimidation law, FCC complaint processes, are triggered after a call happens and resolved on a timeline measured in months. A voter misled by an AI-cloned candidate's voice the weekend before Election Day has no practical recourse before ballots are counted.
There's a sharper irony underneath the timing. This is the same commission that, after a 2024 AI-cloned robocall impersonating President Biden told New Hampshire voters to skip their primary, ruled that AI voice-cloning in robocalls violates the TCPA outright, and that fined conservative operatives Jacob Wohl and Jack Burkman roughly $5.1 million over a 2020 voter-suppression robocall scheme targeting Black neighborhoods. The commission's own precedent treats AI-voiced political calls as a documented vector for election deception. The waiver doesn't repeal that finding; it just carves out an exception for the specific calls most likely to use the technology the commission already flagged as dangerous.
The structural point getting the least attention is the vehicle itself. A waiver lets a 2-1 Republican majority bypass the extended notice-and-comment process a permanent rule change would require, turning what functions as a major shift in wireless privacy protection into something closer to a single commission vote. Commissioner Anna Gomez, the lone Democrat, has called the prospect a 'tsunami of robocalls and misinformation,' but a dissent on a three-member panel doesn't stop a majority ruling. If the waiver is granted this week, it will have been decided by three people, on a schedule designed to outrun both litigation and the election itself.