本頁4 分鐘閱讀
概述
The safer position is using tools under enterprise terms that bar training, limit retention and restrict access, because both the ethical duty of confidentiality and a court's view of waiver depend on whether confidentiality was reasonably protected.
深入探討
Three protections are often blurred together. Attorney-client privilege is an evidentiary rule shielding confidential communications between lawyer and client made to obtain or give legal advice. Work product protects materials prepared in anticipation of litigation. The ethical duty of confidentiality under Rule 1.6 is broader, covering all information relating to the representation, and requires reasonable efforts to prevent unauthorized disclosure. Privilege is generally waived by voluntary disclosure to a third party outside the privileged relationship. Courts have long accepted, however, that lawyers can use service providers such as email hosts, cloud storage and eDiscovery vendors without waiver when confidentiality is reasonably maintained. AI tools are likely to be analyzed similarly, which makes the vendor's terms central. Case law applying waiver doctrine specifically to AI prompts is still thin, so lawyers should not assume a favorable outcome. Risk rises when terms let the provider use inputs to train models, keep them indefinitely, or allow human reviewers to read them, because each undercuts a claim that the communication stayed confidential. Consumer versions of popular chatbots have generally allowed training on conversations by default unless users opt out, while business and API offerings typically do not. Litigation can also affect retention: in 2025, the court in The New York Times's copyright suit against OpenAI ordered the company, for a period, to preserve user chat logs, including ones users had deleted. Work product is harder to waive; disclosure generally waives it only if it substantially increases the chance that an adversary obtains the material. Clients raise a separate issue. When they discuss their legal problems with a chatbot before or instead of talking to counsel, those conversations are not communications with a lawyer and may be discoverable. A common misconception is that turning off chat history makes a consumer tool safe. It may limit training use, but providers often still retain data for a period for abuse monitoring.
戰略影響
風險與安全
災難性和日常的人工智慧危害都取決於誰了解風險以及誰能夠採取行動。
更明確的決策
民眾和專業素養決定強而有力的安全政策在政治上是否可行。
突破炒作
清晰的解釋可以減少炒作、實驗室公關和模糊道德劇場的影響。
The Future of Generative AI and Attorney-Client Privilege
Courts have only begun addressing how privilege and work product apply to AI prompts and outputs, and early decisions may differ by jurisdiction and facts. Commentators and some policymakers have discussed whether conversations with AI should receive special protection, but no such privilege currently exists. Meanwhile, vendors increasingly offer legal-industry terms, private deployments and zero-retention options, which make confidentiality easier to defend. Lawyers should expect more discovery requests aimed at AI chat logs, both their own and their clients', and more firms advising clients directly not to discuss their matters with consumer chatbots.
現實世界的實施
A lawyer pastes a client's email about an internal investigation into a free consumer chatbot with default settings that allow training; even if privilege survives, the lawyer may have fallen short of the duty under Rule 1.6 to take reasonable measures to protect client information.
A firm licenses an AI tool under an enterprise agreement that specifies no training on customer data, a defined retention period, encryption and no vendor access without permission, and records that review in its vendor file.
A client works through the facts of a dispute with a consumer chatbot before calling a lawyer; those chats are communications with a third-party service, not with counsel, and may be discoverable.
Before running a sensitive matter through an AI tool, a lawyer removes client names and identifying details, reducing what would be exposed if the data were ever accessed.
風險與防護欄
將存在風險視為科幻小說，同時能力複合。
混淆了表面產品安全與高度自治下的對準。
只給非英語和非專業觀眾留下低品質的資源。
實施路線圖
單獨的產品危害、誤用和失控/失調風險。
詢問哪些證據會改變您對時間表和嚴重性的看法。
比起行銷主張，更喜歡主要來源和具體評估。
確定一條行動路徑：職業、政策、資金或技能——而不僅僅是意識。
不斷探索
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Generative AI and Attorney-Client Privilege quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
常見問題
What is Generative AI and Attorney-Client Privilege?
Entering client information into a generative AI tool does not automatically waive attorney-client privilege, but it can put confidentiality at risk when the tool's terms let the provider train on, retain or review inputs. The safer position is using tools under enterprise terms that bar training, limit retention and restrict access, because both the ethical duty of confidentiality and a court's view of waiver depend on whether confidentiality was reasonably protected.
Which protection covers all information relating to a representation, not just lawyer-client communications?
Rule 1.6 confidentiality is broader than privilege, which covers confidential communications for legal advice, and work product, which covers litigation materials.
How have courts generally treated lawyers' use of service providers such as email hosts and cloud storage?
Courts have long accepted that using such providers does not waive privilege if confidentiality is reasonably protected, and AI tools are likely to be analyzed similarly.
Which vendor terms does the guide say increase confidentiality and waiver risk?
Each of these undercuts the claim that the communication stayed confidential. The other options are protective measures.
What did a court order OpenAI to do in 2025 in The New York Times's copyright suit?
The preservation order showed that litigation can override a provider's normal deletion practices, which matters for anything placed in a chat.
Under the guide, when does disclosure generally waive work product protection?
Work product is harder to waive than privilege; the usual test asks whether disclosure made it substantially more likely that an adversary would get the material.
繼續學習
相關指南
為此主題精選的更多指南