Enacted Law Verified June 17, 2026

Arkansas AI Legislation in 2026

Arkansas has enacted four AI laws focused on digital likeness protection, AI content ownership, CSAM prevention, and public entity governance.

4 Tracked bills
4 Enacted

The Arkansas Picture

Arkansas has chosen a deliberately light-touch, builder-friendly posture on artificial intelligence, and the centerpiece is Act 927 of 2025. While most states spent 2025 debating algorithmic discrimination and high-risk AI audits, Arkansas did something almost no other state did. It answered the ownership question. Act 927 declares that if you lawfully supply the data and the prompts, you own the model you train and the content you generate. That clarity is a competitive signal. For founders, studios, and enterprises building on generative tools, Arkansas removed a major source of legal ambiguity around who holds rights to AI outputs and trained weights. I view this as the state's signature contribution to American AI policy, and it is the first thing any business should understand before assuming federal copyright doctrine is the only word on the subject.

The rest of Arkansas's 2025 framework is narrow and harm-specific rather than comprehensive. Act 848 governs the state's own house, requiring every public entity to publish an AI use policy and to keep a human in the loop on final decisions. Act 977 and Act 827 are criminal statutes targeting the two AI harms with the clearest political consensus, namely AI-generated child sexual abuse material and nonconsensual sexual deepfakes. Crucially, the one bill that would have imposed broad obligations on private AI developers, SB258, had its AI provisions stripped and then died. So as of June 2026 there is no Arkansas law imposing impact assessments, bias audits, or high-risk AI duties on private companies.

What this means in practice is that the compliance burden in Arkansas is light but specific. Businesses should treat Act 927 as an opportunity and put their contracts in order. Because ownership defaults can be reassigned by agreement, employment agreements, contractor terms, and customer terms of service should explicitly state who owns AI outputs and trained models, rather than relying on the statutory default. The employer-ownership carve-out only applies to work within the scope of employment and under employer direction, so the scope of AI use in job descriptions matters. Any organization touching synthetic sexual imagery, minors, or likeness-based content faces real criminal exposure under Acts 977 and 827, and any vendor selling to Arkansas government must align with the Act 848 human-decision and transparency requirements.

My guidance to leadership is straightforward. Arkansas is a low-friction jurisdiction for deploying and commercializing AI, and Act 927 is a genuine asset for anyone who wants legal certainty over the AI assets they create. The trade-off is that light-touch does not mean no rules. The criminal deepfake and CSAM statutes are enforced, the public-sector policy mandate is in force, and the absence of a comprehensive law means companies should not assume Arkansas will preempt the federal patchwork. Note that 2026 is an Arkansas fiscal session, so no major new substantive AI statutes are expected this year, and the governing law remains the 2025 acts.

Tracked Arkansas AI Bills

Ownership of Generative AI Models and Content

Governance

Became Act 927 on April 21, 2025, effective August 3, 2025. Declares that the person who lawfully supplies prompts owns the generated content and the person who supplies lawful training data owns the trained model, with an employer carve-out.

Effective: August 3, 2025

Key Provisions

  • Prompt provider owns lawfully generated content
  • Training-data provider owns the trained model
  • Employer owns models and content from in-scope employee use
  • No ownership where output infringes pre-existing IP

Public Entity AI Use Policies

Governance

Became Act 848 on April 17, 2025, effective August 3, 2025. Requires every Arkansas public entity to adopt a published AI use policy and keep a human as the final decision-maker over any AI recommendation.

Effective: August 3, 2025

Key Provisions

  • Mandatory AI use policy for public entities
  • Human final decision over automated recommendations
  • Public availability of policies on request

AI-Generated CSAM

CSAMCrime

Became Act 977 on April 22, 2025. Criminalizes computer-generated visual media indistinguishable from a real child in sexually explicit conduct, with exemptions for law enforcement and AI safety testing.

Effective: July 21, 2025

Key Provisions

  • Criminalizes AI-generated CSAM indistinguishable from real
  • Defines computer generated, indistinguishable, adversarial testing
  • Good-faith exemptions for investigation and safety testing

Unlawful Distribution of Sexual Deepfakes

DeepfakesCrime

Became Act 827 in April 2025, effective August 3, 2025. Creates the criminal offense of distributing deepfake visual material of a sexual nature, with criminal and civil penalties.

Effective: August 3, 2025

Key Provisions

  • Crime of distributing nonconsensual sexual deepfakes
  • Graduated misdemeanor and felony penalties
  • Civil penalties for victims

Download The AI Law Brief of All 50 States

A formatted PDF covering every tracked AI bill across all 50 states, with status, effective dates, and key provisions. Useful for compliance teams and board packs.

* You will also be subscribed to my newsletter.

Other states

← All 50 states and the full report