From the SLED-wide archive edition of September 1, 2026
California bill would prohibit delegating legal judgment and require verification and disclosure
California Legislature · State courts, legal services, and alternative dispute resolution · California, United States
- Publisher
- California lawmakers pass bill governing lawyers' use of AI
- Original publication
- September 1, 2026
- Source retrieved
- Not recorded in the historical archive
What happened
Both chambers of the California Legislature approved SB 574 and sent it to the governor. The measure would prohibit lawyers from delegating the practice of law to generative AI, require reasonable verification and correction of outputs and citations, require disclosure of AI use in court submissions, restrict entry of confidential and nonpublic information, and prohibit arbitrators from delegating decisions to AI.
Why it matters
Courts, public defenders, prosecutors, attorneys general, municipal counsel, hearing officers, and administrative adjudicators all perform high-consequence knowledge work. The bill offers a capability-based control pattern that permits assistance while preserving professional responsibility and decisional authority.
Evidence and measured results
Reuters verified legislative passage on September 1 and linked the official bill text. The bill adds specific duties around confidentiality, accuracy, disclosure, and non-delegation and would require the Judicial Council to revisit its AI standard as the technology develops. It had not yet been signed or implemented, so there is no outcome evidence.
Limitations and uncertainty
SB 574 was awaiting gubernatorial action when reported and may change through signature, veto, litigation, or implementation. Some legal experts told Reuters that parts duplicate existing ethical duties, and the record does not show whether the proposed requirements reduce hallucinated filings or confidentiality incidents.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source as summarized in the preserved archive. Enriched 2026-09-05; this does not change the original publication date. Labels below come from the analysis itself.
Sales
Role takeaway
Problem and stakeholders: Government counsel, courts, defenders, prosecutors, arbitrators, records teams, and security leaders need controls for confidential legal work and reliable citations.
- Discovery
- Who verifies assisted research, what information may enter a tool, and who retains authority to file or decide?
- Value hypothesis
- Protected drafting and verification could support staff while making accountability visible.
- Potential engagement
- Assess one research or drafting workflow with legal leadership and test citation review and confidentiality.
- Evidence boundary
- In the archive, SB 574 awaited gubernatorial action; it must not be presented as enacted law or current legal advice. The reporting does not show reduced hallucinated filings, lower costs, or fewer confidentiality incidents from the proposed duties.
Pre-sales engineering
Role takeaway
- Fit
- Limit initial use to assisted research or drafting under professional review.
- Architecture
- Use protected matter workspaces, case permissions, grounded retrieval, citation provenance, and retained review, separated from filing and adjudication.
- Prerequisites
- Approved sources, confidentiality rules, authorized reviewers, and counsel's confirmation of applicable requirements.
- Constraints
- Sealed, victim, witness, health, and financial information may preclude convenient external tools.
- Security
- Validate matter isolation, data-loss controls, retention, vendor data use, and logs.
- Proposed validation
- Test fabricated citations, misleading quotations, cross-matter retrieval, and unauthorized submission using approved data. Reviewers must be able to verify underlying authority; assistants must not independently exercise filing or decision powers. Grounded retrieval alone does not establish legal accuracy.
Delivery
Role takeaway
Work and dependencies: Legal leadership defines assistance, verification, disclosure, and escalation with security, records, and procurement support. Build a review trail and train practitioners to inspect authority rather than fluent output.
- Ownership
- Practitioners retain professional responsibility; technology teams enforce access and workflow controls.
- Governance checkpoints
- Confirm legal status before adopting obligations, approve confidential-data handling, and reassess model or source changes.
- Skills and adoption
- Include actual reviewers so verification remains feasible within casework.
- Proposed acceptance
- Sample outputs have verified citations and accountable sign-off, matter-access tests pass, required disclosure evidence is retained, and incident remediation is rehearsed.
- Risks
- Pending legislation can change, existing duties may overlap, and nominal human review can fail when time or source access is inadequate.
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Legal AI environments need protected matter workspaces, role and case-based access, source-grounded retrieval, citation provenance, immutable review records, disclosure support, and hard separation between research or drafting and the authoritative act of filing or deciding. Agentic tools should not possess unilateral submission or adjudication authority.
Governance
Who approves, reviews and stays accountable for outcomes?
Translate professional duties into acceptable-use rules, mandatory verification workflows, training, disclosure criteria, sanctions or remediation, and periodic review. Procurement should require confidentiality protections, auditability, model-change notice, and the ability to retain evidence of human review.
Security and privacy
What data, permissions and controls need testing?
Prevent confidential, personal, medical, financial, witness, victim, and sealed information from reaching systems without restricted access and enforceable confidentiality. Apply data-loss prevention, matter-level authorization, logging, retention, and incident handling to legal copilots.
The preserved archive analysis covered architecture, governance and security. Not assessed for this record: accessibility and workforce, procurement, operating model.
Publication history
- 2026-09-01SLED-wide archive · Issue 056 resources
Stable resource ID: california-legal-ai-duty-bill