Arizona Supreme Court Allows Judges to Test AI Under Supervision

Arizona chooses cautious AI adoption over outright ban

The Arizona Supreme Court has declined to ban artificial intelligence in courtrooms, rejecting a January 2026 petition that sought to prohibit judges from using generative AI for core judicial functions through 2029. Instead, the court will maintain its current approach of closely supervised testing.

Under existing guidelines, Arizona judges may deploy approved AI tools to address case backlogs, streamline administrative tasks, and improve public access to justice. The court established a clear boundary, however: judges cannot use AI to draft legal rulings or render final case decisions.

The decision was first reported by 13 News in Tucson.

Legal professionals see AI as evolutionary tool

Jaime Ibrahim, Deputy Director for Southern Arizona Legal Aid and a 20-year legal veteran, compared AI adoption to earlier technological shifts in the profession. "We went from books to online" for case law research, she noted, characterizing AI as another tool when used appropriately.

Ibrahim observed that AI has gradually entered legal practice over recent years, with many attorneys already incorporating it ethically into briefs and court filings. She emphasized the importance of learning proper implementation rather than avoiding the technology entirely.

University of Arizona Law Professor Keith Swisher warned that completely blocking AI access could disadvantage courts. He pointed to emerging challenges where AI literacy is becoming essential—such as detecting deepfake evidence or handling cases involving self-represented litigants who use AI-generated legal advice.

Why it matters

Arizona's decision represents a middle path as courts nationwide grapple with AI integration. By allowing supervised experimentation rather than imposing blanket restrictions, the state positions itself to develop practical guidelines based on real-world testing. The approach acknowledges both AI's potential efficiency gains and the critical need for human oversight in judicial decision-making—a balance other jurisdictions will likely study as they craft their own policies.

Human review remains non-negotiable

Both legal experts and the court stressed that AI cannot substitute for human judgment. "Sometimes AI can make mistakes still, or exaggerate certain points, or miss other points," Swisher cautioned, emphasizing that trained human review remains essential even for limited AI applications.

The Arizona Steering Committee on Artificial Intelligence and the Courts continues monitoring the technology's implementation. The committee is scheduled to deliver its next comprehensive status report to the Supreme Court on March 1, 2027.

Details of the court's decision were reported by 13 News.