HB 24-1147 (2024): Candidate Election Deepfake Disclosures
In effectRequired disclaimer, word for word
“This (image/audio/video/multimedia) has been edited and depicts speech or conduct that falsely appears to be authentic or truthful.”
- Citation
- C.R.S. 1-46-101 to 1-46-106 (deepfake rule at 1-46-103); penalties at C.R.S. 1-45-111.5(1.5)(c.5)
- Enacted
- May 24, 2024
- Effective
- July 1, 2024
- When it applies
- 60 days before a primary election or 90 days before a general election
- Media covered
- image, video, audio, or multimedia content substantially created or modified by generative AI that falsely appears authentic or truthful and depicts an individual appearing to say or do something they did not (candidates include incumbents and federal candidates)
- Who it covers
- any person who distributes, disseminates, publishes, broadcasts, transmits, or displays the communication to an audience that includes the candidate's electorate
- Rule
- ban unless disclosed
- Intent required
- knows or has reckless disregard for the fact that the depicted candidate did not say or do what is depicted
- Disclaimer format
- Visual: text no smaller than the largest font size of other text in the communication (or easily readable if there is no other text). Audio: read in a clearly spoken manner in the same pitch, speed, language, and volume as most of the audio, at the beginning and at the end, and if longer than two minutes, also at intervals of not more than one minute. Metadata must include the disclosure, the identity of the tool used to create the deepfake, and the creation date and time. Disclosure must be permanent or hard to remove where technically feasible. Must also meet any Secretary of State rules and FCC size, duration, and placement rules for broadcast or online communications.
- Exemptions
- interactive computer services under 47 U.S.C. 230 for third-party content
- broadcasters, cable and satellite operators airing a deepfake in bona fide news coverage if they acknowledge questions about its authenticity
- broadcasters, cable and satellite operators, and streaming services paid to air the communication
- websites and regularly published newspapers, magazines, and periodicals carrying news and commentary that clearly state the deepfake does not accurately represent the candidate
- satire or parody that depends substantially on a person's physical or verbal impersonation rather than generative AI
- the technology provider used to create the deepfake
- Civil penalty
- Any person may file a complaint with the Secretary of State, adjudicated under C.R.S. 1-45-111.7 (C.R.S. 1-46-104). A hearing officer must impose a civil penalty of at least $100 per violation when no paid promotion is involved, or at least 10 percent of the amount spent to advertise or promote the communication, or a higher amount based on distribution and exposure (C.R.S. 1-45-111.5(1.5)(c.5)).
- Who can sue
- The depicted candidate may sue for injunctive relief, compensatory and punitive damages, attorney fees and costs, and other relief; must prove knowledge or reckless disregard by clear and convincing evidence.
- Injunction available
- Yes
- Notes
- Applies only to communications about candidates (not ballot measures). Because audio is covered, an AI-voiced robocall or voicemail that imitates a candidate in the window needs the spoken disclosure at the start and end. Applies to communications distributed on or after July 1, 2024. The act preserves criminal liability under C.R.S. 1-13-109 (knowingly or recklessly false statements about candidates) for deepfakes with false statements (C.R.S. 1-46-106). No 2025 or 2026 amendment found; a 2026 biometric likeness felony bill (HB26-1049) died in committee. Injunction suits must be heard by the district court at the earliest practical time (C.R.S. 1-46-105).