Can an Oregon lawyer use a chatbot or AI agent for client intake and marketing on their website?

Plain-English summary

The opinion answers "Yes, qualified": Oregon lawyers may use chatbots, including those built on AI agents, to communicate with clients and prospective clients on their websites and elsewhere for marketing and intake. It identifies the rules implicated by that use and the conditions each imposes.

As a threshold matter, the opinion concludes that under RPC 1.1 a lawyer using a chatbot must understand in sufficient detail how it works, including conducting due diligence on the chatbot vendor and any underlying software the vendor relies on. It treats AI agents, software that can autonomously create sub-tasks and act without intervening human input, as posing greater risk than rule-based chatbots and therefore requiring greater supervision.

On supervision (RPC 5.3), the opinion concludes a lawyer is responsible for a chatbot vendor as a nonlawyer assistant and must make reasonable efforts to ensure it performs competently and protects confidentiality. The opinion identifies several specific risks. First, a chatbot that sounds like a lawyer can lead a prospective client to reasonably believe an attorney-client relationship has formed (the existence of the relationship turns on the client's reasonable belief under In re Weidner), so the opinion concludes the website and chatbot should carry clear disclaimers and a responsible lawyer should be notified to review interactions promptly.

Second, under RPC 7.1 the opinion concludes a chatbot must not make false or misleading statements about the lawyer's services, must not give legal advice (which could implicate the unauthorized practice of law), and, per other states it cites, a chatbot that does not identify itself as a chatbot may itself be misleading. Under RPC 7.3(a) the opinion concludes automated solicitation creates a risk of soliciting a person whose impaired state a human would detect but a chatbot would not. Third, on confidentiality (RPC 1.6) and duties to prospective clients (RPC 1.18), the opinion concludes lawyers must guard information a chatbot collects, including by entering prospective clients into a conflicts system, because disqualifying information received through a chatbot can bar adverse representation. Finally, the opinion concludes autonomous actions by an AI agent (such as sending an engagement letter on its own) should be subject to lawyer approval.

In practice

Under this opinion, an Oregon lawyer who deploys a website chatbot for intake or marketing is acting permissibly only if the lawyer understands the tool, supervises the vendor, and builds in safeguards: clear disclaimers that the chatbot is not a lawyer and that no attorney-client relationship is formed, prompt human review of chatbot conversations, controls against false or misleading statements, and a conflicts-intake step for information the chatbot gathers from prospective clients. The opinion treats AI agents capable of acting on their own as requiring lawyer approval before autonomous communications go out.

The opinion makes the analysis depend on the specific use; it notes that rules beyond the ones it discusses may be implicated depending on how a particular chatbot is configured.

Common questions

Q: Can an Oregon lawyer put a chatbot on their law firm website for client intake?

A: Yes, qualified. The opinion concludes chatbots are permissible for intake and marketing if the lawyer is competent in the tool, supervises it under RPC 5.3, and adds safeguards against misleading communications and inadvertent attorney-client relationships.

Q: Does a law firm chatbot have to disclose that it is not a person?

A: The opinion concludes a chatbot that does not identify itself as a chatbot could be a misleading communication under RPC 7.1, citing Florida and New York City Bar opinions, and that websites and chatbots should include clear disclaimers.

Q: Can a chatbot conversation create an attorney-client relationship?

A: Yes, the opinion concludes it can. Because an attorney-client relationship in Oregon turns on the client's reasonable belief, a chatbot that sounds like a lawyer could create one, triggering duties of competence, diligence, and communication, so the opinion concludes a responsible lawyer should promptly review chatbot interactions.

Q: What happens to confidential information a prospective client gives a chatbot?

A: Under RPC 1.18, the opinion concludes disqualifying information received through a chatbot can bar the firm from representing an adverse party, so lawyers should promptly enter prospective clients into their conflict-management system and protect the information under RPC 1.6.

Background and rules framework

The opinion interprets the Oregon Rules of Professional Conduct for chatbot and AI-agent use: RPC 1.1 (competence), RPC 1.18 (duties to prospective clients), RPC 5.3 (supervision of nonlawyer assistance), RPC 7.1 (false or misleading communications), and RPC 7.3(a) (solicitation of a person who cannot exercise reasonable judgment), with RPC 1.6 governing confidentiality. It builds on OSB Formal Opinion 2025-205 (AI tools) and cites opinions from Florida, the New York City Bar, and a joint Pennsylvania/Philadelphia opinion.

Citations and references

Rules of Professional Conduct:

• Oregon RPC 1.1 / Model Rule 1.1 (competence)
• Oregon RPC 1.18 / Model Rule 1.18 (prospective clients)
• Oregon RPC 5.3 / Model Rule 5.3 (nonlawyer assistance)
• Oregon RPC 7.1 / Model Rule 7.1 (false or misleading communications)
• Oregon RPC 7.3(a) / Model Rule 7.3 (solicitation)
• Oregon RPC 1.6 / Model Rule 1.6 (confidentiality)

Cases:

• In re Weidner, 310 Or. 757, 801 P.2d 828 (1990), attorney-client relationship turns on the client's reasonable belief

Other opinions cited:

• OSB Formal Op. 2025-205: artificial intelligence tools
• Florida Bar Ethics Op. 24-1 (2024): generative AI and chatbot intake
• New York City Bar Formal Op. 2024-5: AI in legal practice
• Pennsylvania Bar / Philadelphia Bar Joint Formal Op. 2024-200: generative AI and unauthorized practice

See also

• OSB Ethics Op. 2025-205: AI Tools in Practice
• FL Bar Ethics Op. 24-1: Generative AI in Practice
• ABA Formal Op. 512: Generative AI Tools
• ABA Formal Op. 492: Obligations to Prospective Clients

Source

• Landing page: https://www.osbar.org/ethics/toc.html
• Original PDF: https://www.osbar.org/_docs/ethics/2026-208.pdf