North Carolina Publishes Statewide AI Roadmap With 17 Goals Across Protection, Workforce and Government Use

North Carolina now has a written plan for artificial intelligence, though not yet a law. On July 1, 2026, Governor Josh Stein's AI Leadership Council released the state's first "Statewide AI Strategic Roadmap," published jointly through the Department of Information Technology and the Department of Commerce ^[1][2]. The document lays out 17 goals arranged under three headings — Protect, Prepare and Transform — with eight goals aimed at shielding residents from AI-driven harms, four aimed at schools and worker retraining, and five aimed at using AI to modernize state government itself ^[1][2]. The council that wrote it was created just ten months earlier, in September 2025, when Stein signed Executive Order 24, pulling in representatives from business, government, education and labor to draft it ^[3].

What the Roadmap Actually Says

Every goal in the document comes with an assigned owner, a target date and, where applicable, a funding note — the trappings of a real policy plan ^[1][2]. But the roadmap itself is explicit that it is advisory: it recommends, it does not compel. Turning most of its 17 goals into enforceable rules would require the Republican-led General Assembly to pass new legislation and appropriate new money, something the council is asking lawmakers to take up in the 2027 session ^[4]. A bipartisan AI Caucus already exists inside the legislature and has said publicly that it intends to convert at least some of the council's recommendations into actual policy, which gives the roadmap a plausible, if unproven, path to becoming binding law ^[4].

The Pressure Underneath

Three structural tensions sit beneath the announcement, and none of them will be resolved by the roadmap alone. The first is a federalism collision: states have long treated consumer protection and their own procurement rules as their own turf, but the Trump administration wants a single national AI standard and is actively working to preempt state-level rules, having signed an executive order in December 2025 directing a challenge to state AI laws and standing up a Department of Justice AI Litigation Task Force on January 9, 2026, to pursue exactly that ^[9]. Any binding law North Carolina eventually passes will be written, and possibly litigated, in the shadow of that federal push ^[9].

The second tension is about who actually holds power here. The roadmap was produced by a council the governor created through executive order, but real authority — the ability to write binding law and to appropriate money — sits with the General Assembly, not the governor's office ^[3][4]. That makes the 2027 legislative session, not this July release, the moment that will determine whether any of this changes residents' lives. The third tension is more subtle: the state wants to deploy AI inside its own operations to modernize services even as it promises to protect residents from AI harms, a dual role that puts government's own growing use of AI on citizens' data under exactly the scrutiny that privacy advocates say is currently missing ^[2][10]. Zooming out, North Carolina isn't breaking new ground so much as joining a wave already in motion — dozens of states have passed roughly 100 AI-related measures, and by one industry tracker's count 46 states already regulate synthetic media such as deepfakes ^[8].

How Each Side Sees It

For Stein's administration, the roadmap is framed as a duty: protecting residents from concrete AI harms like scams, deepfakes and privacy violations, while also preparing the workforce and modernizing government services, all under a "trustworthy AI" banner built with input across party lines and industry sectors ^[1][2][3]. The political logic is straightforward for a first-term Democratic governor in a closely divided state — an early, bipartisan-sounding win on a fast-moving issue doubles as a signal that North Carolina is open for tech investment ^[1][6]. Whether that translates into real authority depends entirely on whether the 2027 legislative package survives; if it stalls, the roadmap remains largely symbolic ^[4].

Business groups and federal-preemption advocates counter that a patchwork of differing state AI rules raises compliance costs and burdens interstate commerce, and argue a single national standard beats 50 states each writing their own ^[5][9]. Their concern isn't that AI shouldn't be governed at all, but that new state protections risk becoming red tape that slows adoption and discourages data-center and technology investment — and they have the Trump administration's DOJ task force as a potential enforcement mechanism against any binding North Carolina law that emerges ^[9].

The Republican-led General Assembly and its bipartisan AI Caucus see themselves, not an appointed council, as the body that actually holds the pen and the purse ^[4]. They've signaled openness to "thoughtful, responsible" rules, but strictly on the legislature's own timeline, which functionally makes them the gatekeepers who can adopt, water down or simply ignore the council's recommendations in 2027 ^[4]. Meanwhile, consumer and labor advocates want enforceable protections against deepfakes, fraud, hiring bias and job displacement, while civil-liberties groups such as EPIC warn separately that government's own AI adoption can expand surveillance of residents' data unless the accountability promises in the roadmap turn out to have real teeth rather than remaining rhetorical ^[10].

How the Coverage Split

Outlets covering the release diverged more in emphasis than in facts. The governor's own press release led with benefit language — "protect," "prepare," "trustworthy" — with the caveat that the plan is non-binding and needs 2027 legislation appearing only well below the headline ^[1]. Progressive outlets like NC Newsline emphasized resident protection and worker equity, treating the goals as largely self-evident goods and giving less space to cost or overreach questions ^[6]. Coverage tied to free-market or federal-preemption interests stressed the risk of a costly, legally uncertain state patchwork, though a direct read of one such outlet's own reporting on the council's creation found it notably more descriptive than combative ^[9][12]. Public radio and trade-press coverage came across as the most consistently neutral, foregrounding the plan's advisory status and the 2027 legislative test that other coverage tended to understate ^[4][7].