Law / United States / 
South Dakota
Election Deepfake Disclosure Law (SB 164, 2025)
SDCL §§ 12-26-32 to 12-26-37 (S.L. 2025, ch. 68)
In force since .
An AI sector rules rule binding public and private bodies.
- How the hook was established
- express
- What makes it apply
- Place of effect
- Obligation class
- Prohibition, Disclosure
- Audit expectation
- none
- Setting the rule reaches
- Elections
What it requires
- Do not disseminate, or contract to disseminate, a deepfake, an AI-manipulated or -generated image, audio recording, or video recording depicting a real individual's speech or conduct that did not occur, within 90 days of a South Dakota election, with intent to injure a candidate, unless the item carries the disclosure required by SDCL § 12-26-37.NIST AI RMFMAP 3.3
- If disseminating such content, superimpose (for an image or video) or speak at the beginning and end (for audio) a disclosure reading that the item has been manipulated or generated by artificial intelligence, sized or spoken to be easily seen or heard by an average viewer or listener.NIST AI RMFMEASURE 2.8
- This prohibition does not reach satire or parody, a bona fide news broadcast that acknowledges questions about the item's authenticity, a paid periodical or website of general circulation carrying news and commentary, or an internet, hosting, or connectivity provider acting in a merely technical or automatic capacity.
- Expect civil exposure to the candidate and the depicted individual, in addition to a Class 1 misdemeanor charge, for a violation.
These lines are LexLint's reading of the law's text.
In the frameworks
- NIST AI RMF
- MAP 3.3Targeted application scope is specified and documented based on the system’s capability, established context, and AI system categorization.
- MEASURE 2.8Risks associated with transparency and accountability – as identified in the MAP function – are examined and documented.
- NIST AI 600-1
- GAI-RISK-07Human-AI Configuration
- GAI-RISK-08Information Integrity
- GAI-RISK-11Obscene, Degrading, and/or Abusive Content
- MIT mitigations Draft
- FINOS AIGF
- AIR-PREV-005System Acceptance Testing
- AIR-DET-021Agent Decision Audit and Explainability
- ETSI EN 304 223
- 5.1.2-1Business need and security risks written down and assessed before deciding to build an AI system.
- 5.1.3-2Threats developers cannot resolve are passed to operators and on to end-users.
- 5.1.4-5End-users told which uses of the system are prohibited.
- 5.3.1-1Users told where and how their data is used, accessed and stored, with developers informing operators.
- 5.3.1-2.1That guidance covers proper use, limits and likely failure modes.
Our reading of where this law's duties sit in each framework, and in the NIST AI RMF, line by line beside each duty above. It never says that running a control meets the law. How we read a framework against the law
Identifiers refer to ETSI EN 304 223 V2.1.1 (2025-12), © ETSI 2025, all rights reserved. ETSI owns the standard; no ETSI text (titles or sentences) is reproduced here. The standard is free to download from ETSI.
If you get it wrong
Criminal exposureYes
Private right of actionYes
Criminal exposure note
A violation of SDCL § 12-26-33 is a Class 1 misdemeanor (up to one year in a county jail or a $2,000 fine, or both, under SDCL § 22-6-2).
Penalty structure
Ceiling for the Class 1 misdemeanor criminal charge; separate civil liability under SDCL § 12-26-36 (damages, costs, and attorney fees) is not itself capped.
- Rule
- Fixed only
- As of
- Currency
- USD
- Fixed cap
- 2,000
Who enforces it
Enforcement body
The Attorney General, an injured or likely-to-be-injured candidate, and the individual depicted may each seek injunctive or other equitable relief; the candidate and the individual depicted may also sue for damages; a violation is a Class 1 misdemeanor.
Settledness
- As of
- Open questions
- Does "with the intent to injure a candidate" in SDCL 12-26-33 require an intent to injure the candidate depicted in the deepfake, or is an intent to injure an opposing candidate by disseminating a deepfake of someone else enough?
- Does "within ninety days of an election" in SDCL 12-26-33 reach any election in South Dakota, or only an election in which the candidate the person intends to injure appears on the ballot?
What it makes you log
Logging duty
Sections 12-26-33 through 12-26-37 create the dissemination prohibition, the disclosure format, the exemptions, and the injunctive and civil remedies. None of them requires a log, record, or documentation of the disclosure or of who received it.
- Kind
- None
- As of
What this law does
SB 164 was signed by the Governor on .
This law prohibits any person, with intent to injure a candidate, from disseminating a deepfake (an image, audio recording, or video recording created or manipulated with artificial intelligence or other digital technology that a reasonable person would believe depicts the real speech or conduct of someone who did not engage in it) within 90 days of an election, unless the item carries the disclosure described in SDCL § 12-26-37; a violation is a Class 1 misdemeanor.
The required disclosure must be sized and placed as the statute specifies for each media type. The prohibition exempts satire or parody, a bona fide news broadcast that acknowledges questions about the item's authenticity, a paid periodical or website of general circulation carrying news and commentary, and an internet, hosting, or connectivity provider acting in a merely technical or automatic capacity.
The Attorney General, an injured or likely-to-be-injured candidate, or the individual depicted may seek injunctive or other equitable relief prohibiting the dissemination. A violator is also civilly liable to the candidate and the depicted individual for damages, costs, and attorney fees, proven by clear and convincing evidence.
When LexLint raises it
When your app profile says your app generates audio with AI, generates images with AI or generates video with AI.
Back to the example · Lint your app
Page updated .