Compliance

MATERIALLY DECEPTIVE MEDIA

AI-produced media that falsely depicts real people and is likely believed.

How the laws define this

Regulators uniformly agree that materially deceptive media consists of image, audio, or video falsely depicting an individual engaging in speech or conduct they did not actually perform, which a reasonable viewer or listener would incorrectly believe to be genuine. Key divergences center on whether intent, consent, or format restrictions are embedded directly within the term's definition. New Mexico HB 182 and Hawaii SB 2687 require a specific intent to alter voting behavior, deceive voters, or harm a candidate's reputation or prospects. Hawaii S.B. 2687 / Act 191, SLH 2024 adds that the media must be an advertisement, while an act amending the Campaign Reporting Act specifies publication without the depicted person's consent. Conversely, provisions such as Michigan Election Law §168.932f (Public Act 265 of 2023) and Act 2024-349; HB 172 establish a core definition based strictly on artificial intelligence generation, false depiction, and reasonable perception, without embedding intent or format rules into the baseline term itself.

Synthesised from the 9 statutory definitions below. Each one is quoted in full, with its source.

How jurisdictions differ

What is distinctive in each definition, binding law in force first. Showing 6 of the 9 instruments that define this term; all are listed under Definitions below.

JurisdictionInstrumentStatusWhat is distinctive
Alabama, United StatesAlabama HB 172 - Deepfakes in ElectionsIn ForceAdds an explicit three-part test requiring the media to be produced by artificial intelligence, with a 90-day pre-election window for prohibitions.
New Mexico, United StatesNew Mexico HB 182 - Deepfakes in ElectionsIn ForceIncludes intent to alter voting behavior by misleading electors as a core element, without mentioning reasonable belief.
Michigan, United StatesDistribution of materially deceptive media; prohibition; violation; misdemeanor; penalties; injunctive relief; definitions (Michigan Election Law §168.932f) — Public Act 265 of 2023 (HB 5144)In ForceSpecifies distribution within 90 days of an election triggers misdemeanor prohibition unless disclaimers are used.
Hawaii, United StatesHawaii SB 2687 - Deepfakes RegulationIn ForceRequires distribution with intent to harm a candidate's reputation or electoral prospects or to change voting behavior.
Michigan, United StatesMichigan Public Acts 263-266 of 2023 - AI in Political AdvertisingIn ForceAdds false depiction of appearance alongside speech or conduct to the scope of deception.
Michigan, United StatesMichigan House Bills 5141-5145 - Regulation of Deepfakes and AI in Election CampaignsIn ForceLimits application to manipulated media intended to influence an election, focusing on disclaimer requirements.

What it triggers

  • In Alabama, distribution by individuals and political campaigns within 90 days of an election is prohibited absent required disclaimers or safe harbors.source
  • In New Mexico, distribution in election campaigns requires clear disclaimers, with criminal penalties for non-compliance.source
  • In Michigan, distribution within 90 days of an election is prohibited without disclaimers, subject to misdemeanor penalties.source
  • In Hawaii, reckless distribution in elections is prohibited, establishing penalties and requiring disclaimers.source
  • In Michigan, political advertisers must disclose such media and are prohibited from using it to mislead voters prior to elections.source

Definitions (9)

Materially deceptive media is AI-produced image, audio, or video that (1) falsely depicts a real individual engaging in speech or conduct they did not engage in, (2) is reasonably likely to cause a viewer or listener to believe the depiction is real, and (3) was produced by artificial intelligence. This three-part test defines the regulated content subject to prohibition absent required disclaimers or safe harbors.

Information (video, image, or audio) that is an advertisement, depicts an individual engaging in speech or conduct they did not in fact engage in, would cause a reasonable viewer or listener to believe the depicted speech or conduct occurred, and was created or manipulated using generative adversarial network techniques, artificial intelligence, or other digital technology.

Refers to an image, audio, or video recording created or altered using artificial intelligence to depict a person engaging in speech or conduct that did not actually occur, with the intent to alter voting behavior by misleading electors.

Any image, audio, or video that falsely depicts an individual engaging in speech or conduct in which the depicted individual did not in fact engage, where a reasonable viewer or listener would incorrectly believe the depicted individual engaged in the speech or conduct, and which was produced by artificial intelligence.

'Materially deceptive media' refers to an image, video, or audio that depicts an individual engaging in speech or conduct that did not occur, or was not uttered or performed by the individual. This media must have been created in whole or in part using artificial intelligence or digitally altered or manipulated, and published without the depicted person's consent.

Media, including video, audio, or images, that is generated or substantially altered through artificial intelligence and falsely depicts a depicted individual engaging in speech or conduct in which the individual did not in fact engage, in a manner that would be convincing to a reasonable person, and is distributed with the intent to harm a candidate's reputation or electoral prospects or to change voting behavior by deceiving voters.

Any image, audio, or video where the media falsely depicts an individual engaging in speech, conduct, or appearance in which the individual did not actually engage, and a reasonable viewer or listener would incorrectly believe that the depicted individual engaged in the speech or conduct.

As defined by HB 182, this refers to an image, audio, or video that portrays an identifiable individual engaging in speech or conduct that did not actually happen, and which a reasonable person would perceive as genuine. It is central to the law's scope.

Any image, audio, or video that falsely depicts an individual engaging in speech or conduct in which the individual did not actually engage, and that a reasonable viewer or listener would incorrectly believe depicts the individual's actual speech or conduct. The definition is used to determine when manipulated media must include disclaimers or is prohibited if intended to influence an election.

Not to be confused with

Deceptive and Fraudulent Synthetic Media

This term emphasizes false depictions of real people likely believed as real, while deceptive-and-fraudulent-synthetic media broadly covers synthetic media misleading about appearance, action, or speech.

Materially deceptive content

This term focuses on AI-produced media falsely depicting individuals, whereas materially deceptive content highlights manipulation likely to harm elections.

Questions and answers

What makes media 'materially deceptive' under these laws?

It must falsely depict a real individual in speech or conduct they did not engage in, appear real to a reasonable person, and typically involve AI generation or alteration. Some definitions add intent to influence elections or timing restrictions like 90 days before voting.

When is a disclaimer required for deepfakes in elections?

Disclaimers are required for distribution of materially deceptive media within 90 days of an election in Alabama and Michigan, or in election campaigns in New Mexico and Hawaii. Michigan also mandates disclosures for political advertising.

Does materially deceptive media always need AI involvement?

Most definitions explicitly require AI production, generation, or manipulation, such as in Alabama, Michigan, and Hawaii. New Mexico focuses on creation or alteration using AI to depict non-occurring speech or conduct.