California State Bar Issues 2026 Ethics Guidance on AI Use in Legal Practice, Emphasizing Human Oversight

The California State Bar has released new guidance for legal professionals on the ethical use of artificial intelligence, updating its 2023 recommendations. The core message is clear: AI tools can support legal work, but they do not absolve lawyers of their professional responsibilities. The guidance specifically addresses generative AI (GenAI) and agentic AI, acknowledging their potential to create new content or autonomously perform tasks. This development is significant for practitioners because it formalizes the expectation that lawyers must maintain technological competence. It directly impacts how legal teams will need to vet and integrate AI tools, emphasizing that professional judgment cannot be delegated to an algorithm. The guidance highlights the ongoing risks of AI 'hallucinations' and the potential for agentic AI to operate without continuous human prompting, necessitating robust oversight. This guidance fits within a broader trend of increasing regulatory scrutiny on AI ethics across various sectors. As AI systems become more sophisticated and integrated into critical decision-making processes, regulatory bodies worldwide are moving from aspirational guidelines to enforceable standards. The EU AI Act, for instance, is already in its enforcement phase, setting precedents for transparency, accountability, and human oversight in high-risk AI applications. This reflects a growing societal demand for trustworthy AI, where the 'algorithm decided' is no longer an acceptable explanation for adverse outcomes. In practice, this means legal professionals must proactively engage with AI ethics. Firms should establish clear policies for AI use, including guidelines for generative and agentic AI, and provide regular training to ensure compliance. Lawyers must understand the terms of service and privacy policies of any AI tool they use, especially when handling confidential client information, and obtain informed consent where necessary. Furthermore, billing practices must accurately reflect the actual costs of AI use, distinguishing between overhead and client-specific charges. The emphasis on verifying AI outputs for accuracy and independently confirming legal citations underscores the non-delegable duty of candor to the court. This guidance serves as a blueprint for responsible AI adoption, mitigating both ethical and litigation risks.