California lawmakers are moving to regulate how attorneys use generative artificial intelligence in their legal practice through new legislation introduced in the state senate. The proposed bill, brought forward by California Senate Judiciary Committee Chair Tom Umberg, a Democratic state senator from Santa Ana, California, would formally embed into law many of the guidelines that the State Bar of California’s Standing Committee on Professional Responsibility and Conduct released in 2023. Those earlier guidelines were intended to help lawyers navigate ethical and professional risks associated with using generative artificial intelligence tools to draft briefs and other filings.
The legislation specifically targets lawyers who rely on generative artificial intelligence to prepare legal documents filed in court. Among its central provisions, it requires attorneys to remove confidential personal identifying information from artificial intelligence systems and to verify the accuracy of artificial intelligence produced material. Umberg explained during a hearing that the bill is designed to ensure human accountability, saying that it basically requires a human to be engaged so that, to the extent that there are cases cited and briefs that are filed, a human has to take responsibility for them and has to disclose whether or not artificial intelligence was used in its presentation. The focus is on maintaining professional standards, preventing the misuse of sensitive data and avoiding inaccurate or fabricated citations in court filings.
The bill does not create any new sanctions for attorneys who violate these requirements, according to reporting from Law.com cited in the article. Instead, it relies on existing state law, under which courts can fine attorneys for not ensuring that they file accurate information. By codifying the State Bar’s guidance without adding new penalties, the measure aims to clarify expectations around artificial intelligence use while preserving judges’ current authority to discipline or fine lawyers for inaccurate or irresponsible filings. The proposal underscores a growing effort within California’s legal community to balance the potential efficiencies of generative artificial intelligence with the ethical duties of competence, confidentiality and candor owed to clients and courts.
