In the race to shape the future of autonomous AI, Delaware is taking a deliberate step forward. The state’s AI Commission has greenlit plans for a “sandbox”—a controlled environment to test cutting-edge agentic AI technologies and their regulations. With its long-standing role as a hub for corporate governance (home to countless U.S. corporations), the state is zeroing in on corporate governance as the first sector to explore, setting the stage to redefine how autonomous AI interacts with business operations.
The Delaware AI Commission signed off on a plan to develop a technology sandbox, where new agentic AI technologies and regulations can be explored. The first sector officials are interested in examining is corporate governance.
Delaware is moving forward with the development of a program to test both the technology tools and the regulatory guidelines for using agentic artificial intelligence.
The Delaware Artificial Intelligence Commission has approved a move to create a framework for a “sandbox” for the testing of innovative and novel technologies that use agentic AI, a form of artificial intelligence that includes autonomous decision-making.
“This is a framework that ultimately could be used in a variety of sectors,” state Rep. Krista Griffith, who chairs the state’s AI commission, said during its quarterly meeting Friday.
The first sector the state is interested in exploring is corporate governance, Griffith said, given Delaware’s prominent place in the corporate world as the home to a number of corporations.
Considering agentic AI’s ability to make decisions and take action, the technology could be used to raise money, or buy companies, Charuni Patibanda-Sanchez, Delaware’s secretary of state, said. However, she noted, it probably could not be used to start a limited liability company (LLC) due to the various “human touches” built into the state’s corporate governance policies.
But, Patibanda-Sanchez proposed to the commission, should the state allow an agentic AI tool to incorporate an LCC? What types of efficiencies could this create for businesses and companies, she asked.
“We’ve realized that we have way more questions than any kind of answer,” the secretary of state said to the commission.
The move to develop a sandbox environment to create and test AI tools in Delaware follows lawmakers’ recent passage of a joint legislative resolution, signed by the governor, directing the Artificial Intelligence Commission to collaborate with the secretary of state’s office in developing a sandbox framework for the testing of AI technologies. Griffith was that resolution’s primary sponsor.
“Delaware is taking this proactive stance to shape how they are developed, deployed, and governed,” Patrick Callahan, a commission member, said during its quarterly meeting. “A sandbox creates a controlled environment for AI agents to be tested, observed and approved, with regulators, and the public, involved from the start.”
Delaware is hardly the only state taking a look at how to regulate, or enable, AI. Last year, Colorado passed the Colorado Artificial Intelligence Act to go into effect in 2026; it focuses on the development and deployment of “high risk” AI and its potential to discriminate around the basis of race, disability, gender and other considerations.
A comprehensive AI regulatory bill passed by the California Legislature last year, the Frontier Artificial Intelligence Models Act, was vetoed by the governor on the grounds that it was too broad.
The White House last week released its AI Action Plan, calling for an expansion of data centers, and a light touch on regulating the new technology. Tech CEOs praised the steps taken by the Trump administration for “removing regulatory barriers to innovation, strengthening the domestic AI stack, and enhancing U.S. global AI diplomacy,” according to a statement by Linda Moore, CEO of TechNet, an industry group.
And the recent federal budget reconciliation, the One Big Beautiful Bill Act, contained for a time a 10-year moratorium on states’ ability to regulate AI. That provision ultimately was removed.
Adam Thierer, resident senior fellow for technology and innovation at the R Street Institute, and a former president of the Progress and Freedom Foundation, characterized state action to regulate AI as a “heavy handed, bureaucratic, confusing, European-style technocratic kind of approach to artificial intelligence policy,” during Wednesday’s episode of Broadband Breakfast Live Online, a weekly panel of experts discussing technology issues.
“These laws comprehensively regulate, but in very different ways, artificial intelligence,” Thierer said, reflecting on what he described as nearly 1,100 AI-related bills pending in U.S. statehouses — 240 of those in just four states: California, Colorado, Illinois and New York.
In Dover, Delaware’s capital, leaders praised the steps taken by The First State to set the stage for AI’s use in corporate governance and other areas.
“It’s a great innovative way to make use of our small state in a competitive way,” Molly DiBianca, a commission member, said.
“This is the place for this,” Patibanda-Sanchez said. “It’s innovative, it’s new, it’s a little scary sometimes. But if we don’t do it, it won’t be the best product out there.”
Delaware’s sandbox isn’t just an experiment in technology—it’s a bet that controlled, sector-specific testing can turn “more questions than answers” into actionable rules for the future of AI in business. As the first state to focus agentic AI testing on corporate governance, it’s leveraging its unique position in the corporate world to lead where others are still deliberating. The road ahead will involve grappling with big questions—about “human touches” in regulation, efficiency versus accountability, and how to balance innovation with guardrails. But by choosing to test, learn, and adapt now, Delaware is positioning itself to shape not just its own AI future, but potentially set a template for how autonomous AI integrates into the corporate world at large.
