Washington state has taken a strong position on AI ethics. The Washington State Bar Association issued Advisory Opinion 202401 in June 2024, providing detailed guidance that includes one of the most notable provisions in the country: a prohibition on using consumer AI tools for client-related work. This puts Washington among the more restrictive states on AI use.
The opinion also takes a client-forward stance on billing, explicitly stating that AI efficiency gains should benefit clients through lower fees. For firms operating in Washington, this opinion sets a high compliance bar that requires both technology vetting and billing practice changes.
What the Bar Says
WSBA Advisory Opinion 202401 (June 2024) covers the full spectrum of AI ethics issues. Rule 1.1 (Competence) requires attorneys to understand AI capabilities and limitations before use. Rule 1.6 (Confidentiality) is interpreted strictly — the opinion prohibits use of consumer AI tools (free ChatGPT, Google Bard, etc.) for any client-related work. Only enterprise-grade tools with contractual data protections pass muster. Rule 5.3 (Supervision) requires treating AI like a non-lawyer assistant with mandatory review of all outputs. The opinion also addresses Rule 1.4 (Communication), recommending client disclosure of AI use.
Billing Implications
Washington's opinion takes the strongest pro-client position on billing of any state. It explicitly states that AI efficiency should benefit clients through lower fees. If AI cuts research time from 8 hours to 1 hour, the client should see that savings reflected. Markup of AI tool costs requires transparency — firms cannot quietly add AI subscriptions to overhead and increase rates. The opinion encourages value-based billing for AI-assisted work and requires clear documentation of which tasks involved AI. This effectively ends the practice of billing AI-generated work at full attorney rates.
Confidentiality Rules
Washington's confidentiality guidance is among the strictest in the nation. The opinion prohibits consumer AI tools for client-related work — period. This means free-tier ChatGPT, Gemini, Claude.ai without enterprise agreements, and similar tools are off-limits for anything involving client information. Approved tools must have: contractual data protections, no training on client inputs, clear data retention policies, and deletion capabilities. Attorneys must conduct due diligence on AI vendors and document their evaluation. The standard is not just "reasonable efforts" — it requires affirmative vendor vetting.
What's Still Unclear
The opinion does not define exactly which tools qualify as "consumer" versus "enterprise" — the line between a paid ChatGPT Plus subscription and a true enterprise agreement is not drawn. It doesn't address whether AI use for internal firm management (not client work) falls under the same restrictions. The interaction with Washington's court-level AI disclosure requirements varies by court. Questions about AI use in pro bono work, where enterprise tool budgets may not exist, are not addressed. The opinion also doesn't specify enforcement mechanisms for the consumer AI prohibition.
The Bottom Line: Washington bans consumer AI for client work and requires AI efficiency gains to benefit clients through lower fees — one of the strictest state positions in the country.
AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.