As a former trial lawyer and current CEO of Merlin Search Technologies, a provider of enterprise-grade GenAI solutions for litigation, I need to sound an alarm about a concerning development from Down Under. The Supreme Court of New South Wales (Australia’s largest state court system) has just issued Practice Note SC Gen 23 – a set of sweeping restrictions on artificial intelligence use that could set a dangerous precedent for courts worldwide.

The Practice Note, effective February 2025, essentially handcuffs lawyers and judges from using AI in critical aspects of litigation. Among its most problematic restrictions:

• • • Prohibits using AI tools for drafting or reviewing any witness statements, affidavits, or expert reports – even when the expert witness believes AI would enhance their analysis
    • Requires lawyers to certify that AI wasn’t used in document preparation, creating additional administrative burdens
    • Bans judges from using AI tools to help draft or edit judgments, or even to assist in legal research
    • Mandates extensive disclosures whenever AI is used for any aspect of written submissions, with verification requirements that can’t themselves involve AI tools

You can read the new rules on GenAI use (actually non-use) set forth in the Chief Justice’s  “Practice Note” here.

Having spent years developing secure AI solutions for complex litigation, I see three fundamental flaws in this approach:

1. The rules make no distinction between public AI chatbots and secure, enterprise-grade legal platforms. Enterprise legal AI systems – with their dedicated environments, encryption protocols, and comprehensive audit trails – are purpose-built for the security demands of litigation.
2. These blanket prohibitions ignore the reality of modern litigation. Today’s cases often involve millions of documents across multiple jurisdictions. The complexity of contemporary legal practice demands sophisticated technological assistance – not just for efficiency, but for thoroughness and accuracy.
3. Most troublingly, these restrictions undermine the professional judgment of lawyers and judges. As officers of the court, lawyers already operate under strict ethical obligations. The tools they use to draft and review documents matter far less than their duty to ensure accuracy.

Instead of this restrictive approach, we need a framework that:

• Distinguishes between public AI tools and secure legal platforms 
• Focuses on validating outputs rather than restricting inputs 
• Provides practical guidelines while respecting professional judgment 
• Encourages technological innovation in service of better legal outcomes

While this Practice Note only affects one large Australian state, it could influence courts globally. From my perspective leading a legal technology company, I see daily how properly implemented AI enhances, rather than compromises, the quality of legal work. Trying to regulate away technological progress will only hurt the administration of justice. 

We need to encourage the use of this powerful new technology rather than throw up digital roadblocks. If our goal is to enhance access to justice (e.g. Fed.R. Civ. P. Rule 1), we should celebrate the proper use of these efficient and cost-effective AI tools. Banning GenAI tools, requiring advance permission just to use them, makes about as much sense as going back to quill pens because they’re more “secure” than word processors. The future of law lies in embracing innovation, not restricting it.