US Supreme Court Justice Barrett Confirms Court Is Refusing to Use AI Tools Over Security Concerns, Sharpening Judicial-Commercial Divide on Legal AI Adoption
Justice Amy Coney Barrett publicly confirmed on 10 May 2026 that the US Supreme Court is not using artificial intelligence tools, stating plainly: 'The court is not using AI because it would be insecure.' The statement, made at a public appearance at the Crystal Bridges Museum of American Art in Arkansas, was explicitly intended to reassure audiences that the Court's opinions are not AI-generated. The disclosure is notable for what it reveals about the fault lines in legal AI adoption. While commercial law firms — including several Magic Circle and elite US firms — have been actively deploying AI tools for document review, due diligence, contract analysis, and drafting support, the highest court in the US has concluded that the security risks of AI integration are currently too high for its institutional purposes. The concerns centre on the vulnerability of AI systems to data interception, inference attacks (where third parties attempt to extract information about inputs from model outputs), and the opacity of AI reasoning in high-stakes judicial contexts. For UK legal practitioners and regulators, the Supreme Court's position provides a striking data point in the ongoing debate about AI governance in legal institutions. UK regulators and the judiciary have begun to address AI use in legal practice, with varying levels of formal guidance. The divergence between commercial and judicial adoption trajectories is itself a regulatory and risk management issue that law firms advising courts, government bodies, and regulated entities will need to navigate.
Why this matters
Justice Barrett's statement crystallises a tension that is structurally important for the legal AI market: the institutions with the most sensitive information — courts and government agencies — are the most resistant to AI adoption, while commercial law firms are under client and competitive pressure to deploy AI at scale. For City firms, this creates a two-track advisory environment: one for private sector clients pushing AI adoption, and another for public sector and judicial clients where AI governance and security concerns dominate. The SRA's emerging position on AI competence means that UK law firms cannot treat AI adoption as purely a commercial efficiency question — it has professional conduct and regulatory dimensions that require active governance frameworks.
On the Ground
A trainee working on an AI governance matter would assist with AI governance policy drafting, reviewing the firm's or a client's proposed AI use policies against applicable SRA guidance and UK AI policy developments. They would also help prepare vendor due diligence questionnaires for AI tool providers and contribute to regulatory impact assessment memos assessing the risks of specific AI deployments in legal contexts.
Interview prep
Soundbite
The Supreme Court's AI refusal on security grounds shows that judicial and commercial adoption curves are diverging — creating two distinct regulatory advisory markets.
Question you might get
“What security and professional conduct concerns should a law firm consider before deploying an AI tool to assist with research or drafting in matters involving confidential client information?”
Full answer
Justice Barrett has confirmed the US Supreme Court refuses to use AI tools due to data security concerns, in a direct counterpoint to the rapid AI adoption underway at commercial law firms. This matters because it highlights that the highest-stakes legal institutions — courts, regulators, government bodies — face AI governance challenges that are qualitatively different from those confronting law firms chasing efficiency gains. For UK practitioners, the SRA's emerging AI competence standards and the UK Judiciary's cautious AI guidance signal that professional conduct obligations around AI use are hardening. This suggests the next growth area in legal AI advisory will be governance frameworks for institutions that are resistant to, rather than enthusiastic about, AI deployment.
My notes
saved