Israel Bar Association Bans Autonomous AI From Legal Decisions

Israel's lawyers face strict new AI limits

The National Ethics Committee of the Israel Bar Association has issued comprehensive guidelines prohibiting lawyers from allowing autonomous AI systems to make substantive legal decisions, provide independent advice, draft court documents, or advance proceedings without close human supervision.

In a position paper published Thursday and signed by committee chair Adv. Menachem Moskowitz, the regulatory body drew a clear line between AI as a supervised tool and AI as an independent actor. The rules apply to Israel's approximately 100,000 practicing attorneys.

The committee specifically addressed the emergence of "AI agents"—systems capable of operating with higher autonomy, planning workflows, selecting tools, and interacting with external systems without continuous human intervention. As these systems grow more sophisticated, the ethical and procedural risks escalate proportionally, the committee warned.

What lawyers can and cannot do

The guidelines establish that autonomous systems must not make decisions in place of lawyers, including matters of client representation, legal advice, court filings, document service, or case advancement without sufficient human oversight and approval.

However, the Bar Association permits AI use for technical, circumscribed, documented, and reversible actions—provided they follow a prior human decision, operate within clear parameters, and remain subject to appropriate controls and human oversight.

The committee emphasized that lawyers bear "personal and direct responsibility for every professional action, decision, piece of advice, argument, or output performed by them or on their behalf through the use of artificial intelligence systems."

Client confidentiality takes center stage

The position paper devotes substantial attention to protecting privileged and confidential client information. Lawyers must not enter "restricted information"—any non-public data subject to legal or ethical confidentiality obligations—into open, unsecured AI systems such as public chatbot platforms.

The Bar Association warned that simple anonymization, like removing a client's name, provides insufficient protection. Sophisticated data cross-referencing could still reveal identities.

Even when using "closed platforms"—AI systems with organizational security commitments and contractual protections—lawyers must examine residual risks including data storage methods, access permissions, encryption levels, and system interfaces.

Disclosure and consent requirements

While routine technical AI use doesn't necessarily require client disclosure, transparency becomes mandatory when AI materially affects representation objectives, when restricted information enters open platforms, when systems influence professional decision-making, or when clients request information about AI usage.

Vague provisions in engagement documents stating lawyers "are permitted to use artificial intelligence" do not constitute sufficient informed consent for using client data in AI systems.

The hallucination problem

The committee specifically warned against AI "hallucinations"—fabricated court rulings, citations, or legal sources that appear credible but are entirely false. Lawyers have already submitted documents containing invented case law to courts, the committee noted.

"AI outputs may provide a lawyer with insights, recommendations, lines of reasoning, drafting assistance, or an initial basis for work," the document states. "However, they cannot substitute for the human professional action or judgment required in any legal matter."

Why it matters

As AI agents gain capabilities to act with greater autonomy, legal regulators worldwide are grappling with where to draw boundaries. Israel's approach—allowing supervised AI assistance while prohibiting autonomous decision-making—offers a framework other jurisdictions may examine as they develop their own rules. The guidelines acknowledge AI's utility while insisting that professional judgment, client confidentiality, and legal accountability cannot be delegated to machines. For the legal tech industry, these rules signal that autonomous AI products targeting lawyers will face significant regulatory scrutiny.

The National Ethics Committee warned that failure to comply with these provisions may constitute prima facie evidence that a lawyer has not met ethical obligations, though compliance alone doesn't absolve lawyers of their duty to exercise case-specific judgment.

These details were first reported by Calcalist.