FCC opens inquiry on AI-generated political ads

The Federal Communications Commission opened a formal inquiry this week on the use of AI-generated content in political advertising, citing concerns about deepfakes and voter manipulation.

The inquiry, announced by FCC Chairwoman Jessica Rosenworcel, asks:

  • Should AI-generated political ads be required to include disclosure?
  • What disclosure format is appropriate (text, audio, visual)?
  • Should there be a database of AI-generated political ads?
  • Should there be restrictions on AI-generated political ads during certain windows (e.g., 60 days before election)?

The inquiry is open for public comment for 90 days. After the comment period, the FCC will determine whether to issue a rule.

Why this matters:

The 2024 US election cycle saw the first significant use of AI-generated political content, including:

  • AI-generated robocalls using a cloned President Biden voice (banned by New Hampshire).
  • AI-generated images of Trump being arrested (widely shared).
  • AI-generated video of a fake Biden campaign announcement.

The 2026 midterms will see more AI-generated political content, and the 2028 presidential election will see much more. The FCC’s inquiry is part of a broader push to establish rules before the 2028 cycle.

The legal context:

The FCC has authority over broadcasting (TV, radio) but limited authority over online content. Most political advertising now happens online, where the FCC has no jurisdiction.

The inquiry will likely result in rules for broadcast AI-generated political ads but leave online AI-generated political content to other regulators (FTC, state attorneys general).

Some states have already acted. California, Texas, and Minnesota have laws requiring disclosure of AI-generated political content. The federal inquiry is partly designed to coordinate state laws.

Industry response:

The advertising industry is split. Major broadcasters (NBC, CBS, ABC) have voluntarily committed to disclosing AI-generated political content. Smaller broadcasters and online platforms have been slower to commit.

Political campaigns have also split. Some campaigns have used AI-generated content with full disclosure. Others have used AI-generated content without disclosure, betting that the legal consequences are minimal.

The FCC’s inquiry will likely result in disclosure requirements but probably not outright bans. Banning AI-generated political content raises First Amendment issues that the FCC would rather avoid. Disclosure is more legally defensible and addresses the most concerning use cases.

The deeper issue:

AI-generated political content is part of a broader challenge to democratic discourse. The same generative AI tools that create value in commerce can be used for voter manipulation. The FCC inquiry is one of many efforts to address this.

The most ambitious proposals would require:

  • Watermarking of all AI-generated content (technically feasible but politically difficult).
  • Real-time fact-checking of AI-generated claims (technically possible but politically contested).
  • Liability for AI-generated disinformation (legally complex).

The FCC’s inquiry will likely settle on disclosure requirements as the most politically feasible approach. Whether disclosure is sufficient is the question for the 2028 cycle.