All articles
From 1 January, a California lawyer who uses generative AI on a court document must tell…
From 1 January, a California lawyer who uses generative AI on a court document must tell the court.
Every document. The law doesn't say how much AI use triggers it.
Governor Newsom signed SB 574 on 30 September. It also bars lawyers from delegating the practice of law to AI, requires them to verify citations, and keeps confidential information out of tools that aren't locked down.
Australia has taken a narrower path. The Federal Court's April 2026 practice note asks for disclosure where AI has shaped evidence and, where AI was used, requires the lawyer to personally verify the authorities in submissions. The NSW Supreme Court's practice note has applied since February 2025.
My view: verification protects clients. Blanket disclosure mostly produces boilerplate.
AI now sits inside most research and drafting tools. A line saying "AI was used" tells a judge nothing about whether anyone checked the work.
What matters is who signed it and what they checked. Under the Uniform Law, that has always been the lawyer. In an incorporated legal practice like Awelle, the practice is accountable for the systems behind that sign-off too.
If Australia follows California, I'd want the disclosure to say who verified the work, not which tool touched it.
Should Australian courts require disclosure of every AI use, or is mandatory verification enough?
#AusLaw #LegalEthics #AIRegulation

View on LinkedIn