Maryland finds one violation in five AI election complaints as enforcement limits emerge

A campaign advertisement gave Maryland voters the wrong Election Day date—the only violation officials identified among five complaints under the state’s new AI election law. Courtney Bell’s November 2 reporting for The Baltimore Sun puts a concrete case behind the law while exposing its limits: election officials cannot address every misleading campaign message.

A June law draws a narrow boundary

Enacted in June, the law gives the State Board of Elections limited authority over election disinformation. Board spokesperson Lexy Rhoden said that authority covers false information about an election’s time, place or manner, its results, or voting rights. It is not a general power to police all misleading political content.

That boundary reflects a constitutional challenge, said Donald Tobin, an election law professor at the University of Maryland’s Carey School of Law. Lawmakers had to balance harmful AI use against protections for political speech and artistic expression. Tobin said the law requires an intent to harm someone, and he expects legal challenges over how courts apply that requirement.

Campaign images test the boundary

Weeks after the law took effect, former Republican gubernatorial candidate Ed Hale posted an AI-generated image of his then-primary opponent, Dan Cox, embracing Gov. Wes Moore. Moore spokesperson Carter Elliott said the post was reported to the elections board. The board had not provided an update on that complaint when the Sun published its story.

Moore’s lawyers also reported an advertisement, alleging it used his likeness to falsely suggest he opposed November’s redistricting ballot measure. The board had not answered repeated questions about that complaint’s status. Neither complaint was presented as a confirmed violation in the Sun’s account.

Another episode involved a newspaper image rather than a fabricated candidate scene. Chick and Ruth’s Delly, owned by District 7 Democratic candidate Spencer Jones, posted an AI-altered Capital Gazette front page. It incorrectly called Jones the incumbent and added praise for the restaurant that did not appear in the newspaper’s story.

Jones said an advertising agency had used AI to sharpen a photograph of the page, and the system invented the additions. He apologized, removed the post and issued a corrected version the following day. His explanation attributed the errors to an image-cleanup task, not a deliberate attempt to fabricate newspaper coverage.

Few local reports, unresolved deterrence

The Sun contacted nine local election directors; four responded, and none said they had reported AI-generated election disinformation this season. Cecil County director Gary Holmes had seen AI-generated campaign promotion but nothing he considered election disinformation warranting a report to the state board.

Baltimore County still had misinformation to correct. Election director Ruie Marie Lavoie said a false social media claim that voters needed postage to return mail-in ballots prompted residents to call her office. That claim was not AI-generated.

Violations of the deepfake election law can carry up to five years in prison or a $5,000 fine. But Candace Turitto, director of Maryland’s Applied Political Analytics Program, questioned whether penalties would deter every campaign. In her hypothetical, misleading material could attract coverage or persuade voters before legal action arrives after an election.

For voters, local officials urged checking dates, polling places and voting hours directly with election offices. The state board also accepts reports of suspected false election information whether or not AI created it. Meanwhile, the state prosecutor’s office declined to discuss ongoing AI misinformation investigations or its review process, citing office policy.