WeVote

Bill

Bill

HJR 7

DIRECTING THE ARTIFICIAL INTELLIGENCE COMMISSION TO WORK IN COLLABORATION WITH THE SECRETARY OF STATE TO CREATE A REGULATORY SANDBOX FRAMEWORK FOR THE TESTING OF INNOVATIVE AND NOVEL TECHNOLOGIES THAT UTILIZE AGENTIC ARTIFICIAL INTELLIGENCE.

153rd General Assembly (2025-2026) Introduced by Krista Griffith and 4 co-sponsors

Delaware creates an AI regulatory sandbox allowing companies to test autonomous AI systems with relaxed requirements under state oversight and the Secretary of State.

Signed by Governor
0
WeVote Research Nonpartisan
Bill Summary · HJR 7

Legislative bill overview

HJR 7 directs Delaware's Artificial Intelligence Commission to work with the Secretary of State to establish a regulatory sandbox—a controlled testing environment where companies can develop and pilot innovative AI technologies, particularly those using "agentic" AI (autonomous decision-making systems), with relaxed regulatory requirements. The bill has been signed into law and creates a framework allowing novel AI applications to be tested before full regulatory compliance is required.

Why is this important

Regulatory sandboxes have become tools for jurisdictions seeking to attract tech innovation while managing risk. For Delaware, this positions the state as AI-friendly and potentially attracts companies and investment in emerging AI sectors. However, the outcome depends heavily on how the Commission designs the sandbox's guardrails, oversight mechanisms, and exit criteria—poorly designed sandboxes can inadvertently expose users or the public to untested, potentially harmful AI systems.

Potential points of contention

  • Regulatory capture risk: Companies testing in the sandbox may have disproportionate influence over the rules governing their own technologies, potentially weakening consumer protections or safety standards
  • Accountability gaps: Agentic AI systems that make autonomous decisions raise questions about liability—if a sandboxed AI causes harm, who is responsible? The bill doesn't specify this clearly
  • "Novel technologies" ambiguity: The bill lacks definition of what qualifies for sandbox testing, creating potential for regulatory scope creep or inconsistent application across different AI applications

Compiled from official sources — confirm details with the bill’s official record.

Sign in to ask a question.