New Hampshire AI Legislation in 2026

The New Hampshire Picture

New Hampshire has deliberately declined to adopt a single comprehensive AI law. When a sweeping framework arrived in the form of HB 1725, the Responsible AI Governance Act, the committee rejected it unanimously as inexpedient to legislate. What I see instead is a state that has chosen a narrow, surgical approach: regulate specific harms where they appear rather than impose a Colorado-style or EU-style horizontal regime. For executives, this matters. There is no New Hampshire high-risk AI system classification, no statewide impact-assessment mandate, and no general algorithmic-discrimination duty on private deployers. The compliance surface is defined by four discrete enacted statutes, not by one omnibus code.

The most distinctive feature of New Hampshire's posture is that the state turned the regulatory lens on itself first. HB 1688 restrains government, not industry. It bars state agencies from using AI to manipulate, surveil, or discriminate against the public, prohibits real-time biometric surveillance outside a narrow missing-persons exception, and forces a human into the loop for any irreversible AI-driven decision. This government-first orientation is unusual and reflects New Hampshire's libertarian civic culture. If you sell AI to New Hampshire state agencies, your product must accommodate disclosure, human review, and the surveillance prohibition, because the procuring agency carries the legal obligation.

The second defining thread is the state's rapid, concrete response to the January 2024 AI Biden robocall that urged voters to skip the primary. That single incident generated two enacted laws and a high-profile enforcement record. HB 1432 created the first state private right of action for deepfake victims, and HB 1596 requires AI-use disclosure in political advertising within 90 days of an election. Separately, the original robocall was prosecuted under existing voter-suppression law and federally by the FCC, which levied a $6 million fine against the consultant responsible. The lesson for builders is that New Hampshire treats synthetic media impersonation and election manipulation as serious, litigable harms with real teeth, including damages payable to private plaintiffs.

For builders and businesses, my practical guidance is this. If you operate consumer or companion chatbots reachable by minors, HB 143 is now in force as of January 1, 2026, and it carries statutory damages from $1,000 per violation, a published safety-protocol requirement, and only a 90-day cure window before the Attorney General can act, so build self-harm and harmful-content guardrails and publish your protocol now. If you touch political advertising or any synthetic likeness, assume disclosure obligations and private-suit exposure under HB 1596 and HB 1432. If you are an AI vendor to health insurers, watch HB 1406, which passed committee but is not yet law, and plan for a likely January 2027 effective date. The throughline is that New Hampshire regulates AI by harm and by sector, so map your product to the specific statute that touches it.