Maryland just became the 30th state to police election deepfakes—here’s the loophole that still makes “AI-proof” video basically impossible
Maryland’s new law can punish materially false deepfakes and force fast corrections—but it can’t simply order platforms to take viral posts down. That enforcement gap is the real story.
By TheMurrow Editorial
May 16, 2026
Key Points
- 1Sets real teeth: SB 141 criminalizes materially false election deepfakes, with penalties up to five years in jail and $5,000.
- 2Builds speed: a rapid-response system requires the Administrator to push accurate corrections fast, plus subpoenas and limited injunction options.
- 3Reveals the loophole: Section 230 limits platform takedowns, meaning Maryland can target creators and spreaders—not always the viral distribution itself.
Maryland just became the latest state to legislate against election deepfakes—and it did so with unusual ambition. On May 12, 2026, Gov. Wes Moore approved SB 141/HB 145, a law with the deliberately plain title “Election Law – Election Misinformation, Election Disinformation, and Deepfakes.” In legislative terms, it’s Chapter 444. In political terms, it’s a warning shot: the era of “I never said that” has arrived at ballot-box speed.
Yet the most revealing part of Maryland’s new law isn’t how tough it sounds. It’s the quiet admission built into the fine print: the state can punish a person for creating a materially false deepfake, and it can mobilize a rapid response to correct election lies—but it cannot simply order social media platforms to take the post down the way many voters assume it can.
That gap—between how synthetic media spreads and what state power can realistically compel—is why Maryland’s move matters beyond Annapolis. The law promises accountability, but it also underscores a basic problem: the internet’s distribution machinery was designed to be frictionless, and election law was not.
“Maryland can chase the author of a deepfake. It can’t always stop the deepfake from traveling.”
— — TheMurrow Editorial
Maryland’s new election deepfake law—what actually happened
On May 12, 2026, Maryland enacted SB 141 (crossfiled as HB 145), creating new rules aimed at election misinformation, election disinformation, and deepfakes. The bill’s fiscal/policy note states an effective date of June 1, 2026—a detail that matters because “enacted” and “in effect” aren’t the same thing in the world of election compliance.
Maryland’s timing also feeds a national scoreboard mentality. The National Conference of State Legislatures (NCSL) now reports that “Twenty-nine states have enacted laws regulating the use of deepfakes in political messaging.” That number has become a flashpoint because some commentary has described Maryland as the “30th state” to act. The discrepancy can come from what, exactly, is being counted: deepfakes-only statutes versus broader “synthetic media + election misinfo/disinfo” frameworks, or laws “enacted” versus laws already “in effect.”
Readers should treat “30th state” as a framing—not a settled census—unless the claim is paired with a specific methodology. Maryland is unquestionably part of the trend, and it arrives with a relatively expansive approach. The law doesn’t just target manipulated media; it also builds a state-run response mechanism for election lies and allows the state to pursue cost recovery in civil court.
Maryland’s timing also feeds a national scoreboard mentality. The National Conference of State Legislatures (NCSL) now reports that “Twenty-nine states have enacted laws regulating the use of deepfakes in political messaging.” That number has become a flashpoint because some commentary has described Maryland as the “30th state” to act. The discrepancy can come from what, exactly, is being counted: deepfakes-only statutes versus broader “synthetic media + election misinfo/disinfo” frameworks, or laws “enacted” versus laws already “in effect.”
Readers should treat “30th state” as a framing—not a settled census—unless the claim is paired with a specific methodology. Maryland is unquestionably part of the trend, and it arrives with a relatively expansive approach. The law doesn’t just target manipulated media; it also builds a state-run response mechanism for election lies and allows the state to pursue cost recovery in civil court.
The headline numbers readers should know
- May 12, 2026: Gov. Wes Moore approved SB 141 (Chapter 444).
- June 1, 2026: The law takes effect, per Maryland’s fiscal/policy note.
- 29 states: NCSL’s count of states with enacted laws regulating deepfakes in political messaging.
- Up to 5 years in jail and/or $5,000 fine: Maximum criminal penalty under Maryland’s deepfake prohibition.
- June 1, 2026: The law takes effect, per Maryland’s fiscal/policy note.
- 29 states: NCSL’s count of states with enacted laws regulating deepfakes in political messaging.
- Up to 5 years in jail and/or $5,000 fine: Maximum criminal penalty under Maryland’s deepfake prohibition.
May 12, 2026
Gov. Wes Moore approved SB 141 (Chapter 444), Maryland’s new election misinformation/disinformation and deepfakes law.
June 1, 2026
The fiscal/policy note states this is when SB 141 takes effect—separate from the date it was enacted.
29 states
NCSL’s reported number of states that have enacted laws regulating the use of deepfakes in political messaging.
5 years / $5,000
Maximum criminal penalty in Maryland for certain election-related deepfake conduct under SB 141.
What SB 141 criminalizes—and what it doesn’t
Maryland’s most forceful tool is a new criminal prohibition aimed at “materially false” deepfakes used to interfere with elections. The law targets conduct that is knowing or done with reckless disregard—a key standard, because election content is often spread by people who did not create it, did not verify it, and did not intend harm.
Under the fiscal note’s description, SB 141 prohibits creating, using, or disseminating a deepfake to produce materially false information when the intent is to do things like:
- influence whether or how a person votes,
- misrepresent facts about registration, voting methods, election results, election processes, or election officials’ duties,
- influence petition-signing for ballot questions,
- and the conduct results in, or is intended to result in, specified harms.
The penalty is not symbolic. The law makes the offense a misdemeanor punishable by up to $5,000 and/or up to 5 years imprisonment. For political consultants, campaign staffers, and independent spenders, that’s a serious risk calculus.
At the same time, the law is not a blanket ban on political satire, sharp editing, or campaign hard-hitting. It focuses on deepfakes that create materially false information and are tied to election interference. That “materially false” threshold is doing important work; it signals that the state is trying—at least on paper—to avoid policing ordinary political speech.
Under the fiscal note’s description, SB 141 prohibits creating, using, or disseminating a deepfake to produce materially false information when the intent is to do things like:
- influence whether or how a person votes,
- misrepresent facts about registration, voting methods, election results, election processes, or election officials’ duties,
- influence petition-signing for ballot questions,
- and the conduct results in, or is intended to result in, specified harms.
The penalty is not symbolic. The law makes the offense a misdemeanor punishable by up to $5,000 and/or up to 5 years imprisonment. For political consultants, campaign staffers, and independent spenders, that’s a serious risk calculus.
At the same time, the law is not a blanket ban on political satire, sharp editing, or campaign hard-hitting. It focuses on deepfakes that create materially false information and are tied to election interference. That “materially false” threshold is doing important work; it signals that the state is trying—at least on paper—to avoid policing ordinary political speech.
“The law doesn’t outlaw persuasion. It targets deception that masquerades as reality.”
— — TheMurrow Editorial
How Maryland defines “deepfake”
SB 141 defines a deepfake as an image, audio recording, or video that is intentionally created or manipulated using generative AI or other digital technology to create a realistic but false depiction—one that an ordinary person would believe is real.
Two elements matter here. First, the law is technology-inclusive: it’s not limited to one model or tool. Second, it uses an “ordinary person” lens, which is both intuitive and contestable. What an ordinary person believes can depend on context, the viewer’s familiarity with politics, and the platform where the clip appears.
Two elements matter here. First, the law is technology-inclusive: it’s not limited to one model or tool. Second, it uses an “ordinary person” lens, which is both intuitive and contestable. What an ordinary person believes can depend on context, the viewer’s familiarity with politics, and the platform where the clip appears.
The “rapid response” system: Maryland’s bet on speed over purity
The most novel part of Maryland’s framework may be administrative rather than criminal. SB 141 builds a kind of rapid response system housed within the election apparatus.
If the State Administrator of Elections receives a credible report that election misinformation or disinformation (including deepfakes) “has been or is being” disseminated, the Administrator:
- must distribute accurate information to correct it,
- may seek an injunction for removal from an online platform (with significant limits),
- may issue subpoenas—consistent with the Stored Communications Act, if applicable—for records about dissemination and targeting.
That “must distribute accurate information” requirement is an important design choice. The law assumes that the first, fastest countermeasure is not prosecution—it’s correction. That approach reflects how modern election harms typically occur: not through a single lie, but through a lie that gains a head start.
If the State Administrator of Elections receives a credible report that election misinformation or disinformation (including deepfakes) “has been or is being” disseminated, the Administrator:
- must distribute accurate information to correct it,
- may seek an injunction for removal from an online platform (with significant limits),
- may issue subpoenas—consistent with the Stored Communications Act, if applicable—for records about dissemination and targeting.
That “must distribute accurate information” requirement is an important design choice. The law assumes that the first, fastest countermeasure is not prosecution—it’s correction. That approach reflects how modern election harms typically occur: not through a single lie, but through a lie that gains a head start.
Misinformation vs. disinformation—Maryland draws the line
Maryland’s fiscal note distinguishes:
- Election misinformation: incorrect or misleading information about the time, place, or manner of an election; results; or voting rights.
- Election disinformation: the same type of content, but knowingly and deliberately disseminated.
That distinction is not merely semantic. It suggests the state intends to differentiate between clumsy amplification and deliberate sabotage—while still granting itself tools to respond to both when the harm is immediate.
- Election misinformation: incorrect or misleading information about the time, place, or manner of an election; results; or voting rights.
- Election disinformation: the same type of content, but knowingly and deliberately disseminated.
That distinction is not merely semantic. It suggests the state intends to differentiate between clumsy amplification and deliberate sabotage—while still granting itself tools to respond to both when the harm is immediate.
The cost-recovery civil suit: a quieter lever with real teeth
Criminal cases are hard. They take time, require proof beyond a reasonable doubt, and can run headlong into constitutional speech challenges. Maryland’s SB 141 adds a different tool: civil cost recovery for the state’s cleanup efforts.
The State Board of Elections, working with the Administrator, may bring a civil action against a person, campaign, PAC, or other entity if the Board finds the defendant created or disseminated election disinformation—or disseminated misinformation with reckless disregard—and the Board incurred costs correcting it.
Courts can award damages and fees only with clear and convincing evidence, a higher standard than ordinary civil cases. The structure signals two priorities: deterrence, and protection against flimsy claims. “Clear and convincing” is meant to make the remedy harder to abuse, while still easier than criminal prosecution.
For campaigns, this is not an abstract concern. Rapid-response corrections cost staff time, paid media, translation, hotline operations, and emergency voter outreach. SB 141 is designed to shift those costs back toward the source of the disruption—at least when the evidence is strong.
The State Board of Elections, working with the Administrator, may bring a civil action against a person, campaign, PAC, or other entity if the Board finds the defendant created or disseminated election disinformation—or disseminated misinformation with reckless disregard—and the Board incurred costs correcting it.
Courts can award damages and fees only with clear and convincing evidence, a higher standard than ordinary civil cases. The structure signals two priorities: deterrence, and protection against flimsy claims. “Clear and convincing” is meant to make the remedy harder to abuse, while still easier than criminal prosecution.
For campaigns, this is not an abstract concern. Rapid-response corrections cost staff time, paid media, translation, hotline operations, and emergency voter outreach. SB 141 is designed to shift those costs back toward the source of the disruption—at least when the evidence is strong.
“The most expensive part of election sabotage is often the cleanup—and Maryland wants the saboteur to pay.”
— — TheMurrow Editorial
Practical takeaway for campaigns and advocacy groups
Maryland’s civil hook means compliance can’t stop at “don’t create deepfakes.” Groups should document:
- sourcing and verification steps for election-related claims,
- internal review processes for ads and mailers,
- vendor contracts that prohibit deceptive synthetic media,
- and rapid correction protocols when errors are discovered.
The goal is not moral perfection. The goal is provable diligence when the state starts asking questions.
- sourcing and verification steps for election-related claims,
- internal review processes for ads and mailers,
- vendor contracts that prohibit deceptive synthetic media,
- and rapid correction protocols when errors are discovered.
The goal is not moral perfection. The goal is provable diligence when the state starts asking questions.
Compliance documentation Maryland’s civil hook incentivizes
- ✓Sourcing and verification steps for election-related claims
- ✓Internal review processes for ads and mailers
- ✓Vendor contracts that prohibit deceptive synthetic media
- ✓Rapid correction protocols when errors are discovered
The loophole people miss: platforms, takedowns, and Section 230 reality
Most voters assume the government can order a platform to remove a viral lie. SB 141 shows how constrained that assumption is—especially at the state level.
The law explicitly carves out immunity for online publishers and service providers where they would have protection under 47 U.S.C. § 230. The fiscal note also states that the State Administrator may not seek an injunction against an “interactive computer service” (as defined in Section 230) for content disseminated by another person through that service.
That sentence contains the real-world limit. Maryland can target the person who made or knowingly spread the deepfake. But if the content is moving through a platform as third-party speech, the state’s takedown power is not straightforward.
The law explicitly carves out immunity for online publishers and service providers where they would have protection under 47 U.S.C. § 230. The fiscal note also states that the State Administrator may not seek an injunction against an “interactive computer service” (as defined in Section 230) for content disseminated by another person through that service.
That sentence contains the real-world limit. Maryland can target the person who made or knowingly spread the deepfake. But if the content is moving through a platform as third-party speech, the state’s takedown power is not straightforward.
Why this matters in practice
A deepfake campaign doesn’t need a single megaphone. It needs many small ones: reposts, group chats, stitched clips, mirrored uploads, and algorithmic recommendations. Even a successful prosecution can arrive after the election—after the harm has been done.
Maryland’s answer is speed: correct quickly, document aggressively, and pursue the actors when evidence supports it. The law recognizes, implicitly, that “removal” is often less realistic than “containment.”
Maryland’s answer is speed: correct quickly, document aggressively, and pursue the actors when evidence supports it. The law recognizes, implicitly, that “removal” is often less realistic than “containment.”
Key Insight
SB 141 is built for containment: rapid correction and targeted accountability—because platform-wide takedown authority is limited by Section 230 realities.
“AI-proof video” and the provenance problem Maryland can’t legislate away
Even if Maryland could surgically remove every viral deepfake, a separate problem remains: verification at scale. The public increasingly has to decide whether to trust video and audio at all.
SB 141’s definition of deepfake—realistic but false depiction that an ordinary person would think is real—captures the threat. But it also highlights the provenance problem: ordinary viewers do not have tools to check where a clip originated, whether it’s been edited, or how it was generated. The law can punish after the fact; it cannot hand every voter a forensic lab.
SB 141’s definition of deepfake—realistic but false depiction that an ordinary person would think is real—captures the threat. But it also highlights the provenance problem: ordinary viewers do not have tools to check where a clip originated, whether it’s been edited, or how it was generated. The law can punish after the fact; it cannot hand every voter a forensic lab.
Case study: the “polling place changed” post
Consider the most common election misinformation pattern: a post that claims a polling place changed, the hours shifted, or a deadline moved. Maryland’s law squarely treats those claims as election misinformation/disinformation when false.
The rapid-response system can push accurate information quickly. The civil tool can recover costs if the misinformation was spread deliberately or recklessly. The criminal provision can apply if a deepfake is used to manufacture “evidence”—a fake recording of an election official giving false instructions, for example.
But the initial advantage often belongs to the lie. The first version travels farthest; corrections chase it. That’s not a Maryland problem. It’s a modern information problem.
The rapid-response system can push accurate information quickly. The civil tool can recover costs if the misinformation was spread deliberately or recklessly. The criminal provision can apply if a deepfake is used to manufacture “evidence”—a fake recording of an election official giving false instructions, for example.
But the initial advantage often belongs to the lie. The first version travels farthest; corrections chase it. That’s not a Maryland problem. It’s a modern information problem.
What readers can do now
Practical steps for voters and civic groups are unglamorous but effective:
- Treat any viral election “alert” as unverified until confirmed by official election channels.
- Save the original post, URL, and screenshots before reporting; evidence disappears quickly.
- Look for specific details: county, precinct, date, and named source—vague claims are a red flag.
- Share corrections with the same intensity as the rumor, especially in private group chats.
- Treat any viral election “alert” as unverified until confirmed by official election channels.
- Save the original post, URL, and screenshots before reporting; evidence disappears quickly.
- Look for specific details: county, precinct, date, and named source—vague claims are a red flag.
- Share corrections with the same intensity as the rumor, especially in private group chats.
Quick verification habits for viral election “alerts”
- ✓Treat viral election alerts as unverified until confirmed by official channels
- ✓Save the original post, URL, and screenshots before reporting
- ✓Look for county, precinct, date, and a named source; vagueness is a red flag
- ✓Share corrections with the same intensity as the rumor, especially in group chats
Free speech concerns—and why Maryland chose a narrow path (mostly)
Any law touching political speech will trigger a legitimate question: does it chill expression? SB 141’s drafters appear to anticipate that challenge by focusing on material falsity, intent (knowing or reckless disregard), and defined election-related harms.
Still, speech advocates can reasonably worry about edge cases. “Reckless disregard” is a familiar legal concept, but it can be contested in practice. What counts as reckless when a campaign shares a clip that “looks real” and fits an existing narrative? What diligence is enough? How much is “ordinary person” perception shaped by partisan expectation?
Maryland’s approach tries to avoid becoming a truth ministry by placing much of the immediate response burden on correction, not censorship. The Administrator “must” distribute accurate information when credible reports emerge. The injunction route exists, but the Section 230 limitation narrows its target set.
The other perspective—often voiced by election administrators and voters—is that the status quo is already chilling, in a different way. Voters who cannot trust information about when and where to vote are effectively deprived of the franchise. SB 141 is built around that democratic priority: elections require shared facts about process even when citizens disagree about outcomes and ideology.
Still, speech advocates can reasonably worry about edge cases. “Reckless disregard” is a familiar legal concept, but it can be contested in practice. What counts as reckless when a campaign shares a clip that “looks real” and fits an existing narrative? What diligence is enough? How much is “ordinary person” perception shaped by partisan expectation?
Maryland’s approach tries to avoid becoming a truth ministry by placing much of the immediate response burden on correction, not censorship. The Administrator “must” distribute accurate information when credible reports emerge. The injunction route exists, but the Section 230 limitation narrows its target set.
The other perspective—often voiced by election administrators and voters—is that the status quo is already chilling, in a different way. Voters who cannot trust information about when and where to vote are effectively deprived of the franchise. SB 141 is built around that democratic priority: elections require shared facts about process even when citizens disagree about outcomes and ideology.
Editor’s Note
SB 141’s design leans on correction over censorship: the Administrator must distribute accurate information once credible reports emerge, while takedown tools are narrower.
Is Maryland really the “30th state”? The counting problem tells its own story
Maryland’s law arrives amid a fast-moving patchwork of state rules on synthetic political media. NCSL currently says 29 states have enacted laws regulating deepfakes in political messaging. Some coverage and commentary label Maryland the “30th,” but that phrasing depends on what is included:
- deepfakes-only statutes vs. broader election misinfo/disinfo regimes,
- “enacted” vs. “effective” dates,
- and differences between trackers.
The more revealing point is not the ordinal number. It’s the direction of travel. States are building a parallel body of election law that assumes manipulated audio and video will be routine—not exceptional.
Maryland’s version is notable because it combines three levers in one package:
1. Criminal penalties for materially false deepfakes used to interfere with voting.
2. An administrative rapid response obligation to correct credible misinformation.
3. A civil cost-recovery pathway when the state pays to clean up the mess.
If you’re looking for a single takeaway, it’s this: Maryland is not merely banning a tool. It’s trying to build institutional muscle for the next information crisis.
- deepfakes-only statutes vs. broader election misinfo/disinfo regimes,
- “enacted” vs. “effective” dates,
- and differences between trackers.
The more revealing point is not the ordinal number. It’s the direction of travel. States are building a parallel body of election law that assumes manipulated audio and video will be routine—not exceptional.
Maryland’s version is notable because it combines three levers in one package:
1. Criminal penalties for materially false deepfakes used to interfere with voting.
2. An administrative rapid response obligation to correct credible misinformation.
3. A civil cost-recovery pathway when the state pays to clean up the mess.
If you’re looking for a single takeaway, it’s this: Maryland is not merely banning a tool. It’s trying to build institutional muscle for the next information crisis.
Maryland’s three-lever framework in SB 141
- 1.Criminal penalties for materially false deepfakes used to interfere with voting
- 2.Administrative rapid response obligation to correct credible misinformation
- 3.Civil cost-recovery pathway when the state pays to clean up the mess
Conclusion: the law’s promise—and the reality check it contains
SB 141 is a serious attempt to meet a serious threat. It recognizes that synthetic media is no longer a niche prank; it’s a scalable tactic for confusing voters about the basic mechanics of democracy. By setting criminal penalties up to five years and building a rapid correction system, Maryland is telling bad actors that “cheap and fast” may no longer be risk-free.
The law also admits, indirectly, what no state can fully solve alone. Section 230 and platform design limit how quickly governments can force the removal of third-party content. Verification remains hard for ordinary viewers. Corrections still lag the first viral lie.
Maryland’s bet is that speed, documentation, and targeted accountability can narrow that gap. Whether it succeeds will depend less on what’s written in Chapter 444 than on what happens the next time a false clip hits the feed at 9 p.m. on the eve of Election Day.
The law also admits, indirectly, what no state can fully solve alone. Section 230 and platform design limit how quickly governments can force the removal of third-party content. Verification remains hard for ordinary viewers. Corrections still lag the first viral lie.
Maryland’s bet is that speed, documentation, and targeted accountability can narrow that gap. Whether it succeeds will depend less on what’s written in Chapter 444 than on what happens the next time a false clip hits the feed at 9 p.m. on the eve of Election Day.
“Removal is often less realistic than containment.”
— — TheMurrow Editorial
T
About the Author
TheMurrow Editorial is a writer for TheMurrow covering explainers.
Frequently Asked Questions
When does Maryland’s SB 141 take effect?
Maryland’s fiscal/policy note for SB 141 states the law takes effect June 1, 2026. The bill was approved by Gov. Wes Moore on May 12, 2026 and is listed as Chapter 444. The difference matters: a law can be enacted before it becomes enforceable.
What is a “deepfake” under Maryland’s law?
SB 141 defines a deepfake as an image, audio recording, or video intentionally created or manipulated using generative AI or other digital technology to produce a realistic but false depiction—one an ordinary person would believe is real. The definition is meant to capture convincing synthetic media, not obvious parody.
What are the criminal penalties for election deepfakes in Maryland?
Under SB 141, certain election-related deepfake conduct is a misdemeanor punishable by up to $5,000 and/or up to 5 years imprisonment. The prohibition targets deepfakes used to create materially false information with knowledge or reckless disregard, tied to influencing voting or election processes.
Can Maryland force social media platforms to take down a deepfake?
Not in the simple, sweeping way many people assume. SB 141 includes a Section 230-related carveout: the deepfake prohibition does not apply to online publishers/service providers where they would have immunity under 47 U.S.C. § 230. The fiscal note also says the Administrator may not seek an injunction against an “interactive computer service” for third-party content.
What is the “rapid response” system in SB 141?
If the State Administrator of Elections receives a credible report that election misinformation/disinformation (including deepfakes) is being disseminated, the Administrator must distribute accurate information to correct it. The Administrator may seek an injunction in limited circumstances and may issue subpoenas consistent with the Stored Communications Act, if applicable.
Why do some people call Maryland the “30th state” on deepfakes if NCSL says 29?
NCSL states that 29 states have enacted laws regulating deepfakes in political messaging. The “30th state” framing can depend on what a tracker counts—deepfakes-only laws versus broader election misinfo/disinfo statutes, or enacted versus effective dates. Without a clearly stated methodology, the safest claim is that Maryland is among a growing group of states legislating in this area.
