AI Privacy Tracker · Version 110426

Lexis+ with Protégé

Plain-English resources for lawyers, judges, and government attorneys who use AI, or want to.

Lexis+ with Protégé · LexisNexis

Lexis+ with Protégé at a glance

Rated Yellow

Usable for client confidences only after you deal with the gap named in the entry.

When you delete a chat

A user can delete their own prompt history. LexisNexis's French-edition help page says a deleted conversation cannot be recovered. LexisNexis does not say how fast deleted data leaves its servers and backups.

PlanRatingTrains on your data?How long keptCan the provider read it?
Lexis+ with ProtégéCustom quote; LexisNexis publishes no pricesYellowNo (contract)Older pages: 90 days; users can deleteSupport staff, pseudonymized (older document)
Protégé General AIPart of a Lexis+ with Protégé subscription (custom quote)YellowNo (contract)As Lexis+; model companies keep none (trust center)Support staff, pseudonymized (older document)

Each plan has one color, set by its contract. The settings in each entry lower the risk but do not change the color. Select a plan for its full entry.

Deleting a chat

What deleting does on Lexis+ with Protégé

Lexis+ with Protégé conversations: what happens after a chat is deleted
LexisNexis's help page (French edition) says a deleted conversation cannot be recovered. The General Terms say each user controls prompt history and its deletion. That control is subject to LexisNexis's automatic deletion controls, retention schedules, and applicable law. LexisNexis does not say how fast deleted data leaves its servers and backups. A delete-all option is not documented. Provider page
Copies kept anyway
  • The General Terms make a user's deletion subject to LexisNexis retention schedules and applicable law. LexisNexis does not publish those schedules.
  • LexisNexis says on its trust center that its model providers keep no prompts, under a zero-retention contract. This is not in the contract.

In short: deleting removes a chat from view, but it does not undo what has already happened to it. On business plans, the organization's retention settings decide how long chats are kept; each plan below explains how.

Plan by plan

Every Lexis+ with Protégé plan in detail

Open a plan for its rating, the settings that lower the risk, and short quotations from LexisNexis's own policy with links.

Lexis+ with ProtégéYellowRated Yellow. LexisNexis's General Terms promise that it will not use your prompts and uploaded files to train large language models, and require it to treat them as confidential. A Data Processing Addendum makes LexisNexis your processor for personal data. What the plan lacks is organization control of retention. Section 4.9 leaves prompt history and its deletion to each user, subject to LexisNexis's own automatic deletion, and the current U.S. terms give no period. LexisNexis also says it removes identifying details from AI interactions so that it can improve performance. The contract does not address that use.

Lexis+ with Protégé (formerly Lexis+ AI), Protégé Legal AI. Your firm, school, legal department, or government office buys a Lexis+ with Protégé subscription. Lexis+ AI was renamed Lexis+ with Protégé in February 2026, so former Lexis+ AI users are on this plan. Price: Custom quote; LexisNexis publishes no prices.

Default settings Yellow

The General Terms forbid LexisNexis to train large language models on your prompts and files and require confidential treatment. A Data Processing Addendum makes LexisNexis your processor for personal data. Each user can delete their own prompt history; the firm has no published retention setting. LexisNexis's French-edition help page and older documents say conversations are deleted automatically after 90 days. The current U.S. pages give no period.

Settings that lower the risk

  1. Personalization settingsWhat it does: LexisNexis says users can adjust the level of personalization or opt out entirely. It says customers decide whether personalization is enabled.
  2. Opt-out of AI failover (General Terms §4.6)What it does: If LexisNexis's cloud provider fails, LexisNexis may move AI processing to a backup location, which may be in another country. The firm can opt out by giving 30 days' written notice. A firm that opts out waives claims for the resulting loss of service.
  3. Protégé BYOK (Bring Your Own Key)What it does: Enterprise and government customers can encrypt stored data with keys held in their own key management service. LexisNexis says the customer can revoke access when required.
  4. Administrator controls for user access and tasksWhat it does: An older LexisNexis document (2023, revised November 2024, Australian site) says administrators can limit which users have the AI. It says they can also switch off document upload and drafting for the whole firm.

Worth knowing

  • The contract's no-training promise covers large language models. The product page says LexisNexis removes identifying details from AI interactions to improve performance. The contract does not rule that use out.
  • Vaults are kept until someone deletes them.
  • Unless the firm opts out by 30 days' written notice, AI processing may fail over to a backup location in another country.
  • The Lexis terms restrict putting Lexis research into outside AI tools. Only tools with zero data retention and a no-training promise qualify.

What the policy says

On no training (binding)
“will not be used by LN to train any large language model(s)”

LexisNexis General Terms and Conditions, §4.3(b), Effective October 19, 2026. Read the source

On confidential treatment (binding)
“shall be treated with at least the same standard of care as required of confidential information under Section 12.1”

LexisNexis General Terms and Conditions, §4.3(b)(ii); §12.1(d), Effective October 19, 2026. Read the source

On processor role
“process the personal data only on your documented instructions”

LexisNexis Data Processing Addendum, §2.3 (incorporated by General Terms §10.3), Last updated October 20, 2026. Read the source

On user controls prompt history (binding)
“The retention of Prompt history and its deletion is within the control of the Authorized User”

LexisNexis General Terms and Conditions, §4.9 Prompt Retention, Effective October 19, 2026. Read the source

On de-identified use to improve performance (product page)
“Identifiable information is removed from AI interactions so that we can improve performance without compromising privacy.”

Lexis+ with Protégé product page, Responsibly developed AI, Accessed October 25, 2026. Read the source

On model-provider zero retention (trust center, not contract)
“we have a zero-retention contract with our foundation providers, ensuring they do not retain any prompts or other data for review and analysis.”

LexisNexis Trust Center, Knowledge Base (FAQ): LLM providers and human reviewers, Accessed October 25, 2026. Read the source

On no sale of personal information
“LN is prohibited from selling or sharing personal information it collects”

U.S. Privacy Laws Addendum, CCPA terms, Last updated January 1, 2026. Read the source

On session purge of uploads; Vaults; 90-day Vault results
“Uploads up to 10 documents, not saved to a Vault, are automatically purged at the end of each session.”

Lexis+ with Protégé product page, FAQ: What are the capabilities of the Vault in Lexis+ with Protégé?, Accessed October 25, 2026. Read the source

On customers decide on personalization
“Customers decide whether personalization services are enabled and whether their organization documents are used in AI responses.”

LexisNexis Protégé product page, FAQ: How does LexisNexis address AI data privacy and security?, Accessed October 25, 2026. Read the source

On customer-managed encryption keys
“to encrypt all customer data at rest using keys that only the customer controls”

LexisNexis Launches Next Evolution of Lexis+ with Protégé (press release), Protégé BYOK, May 7, 2026. Read the source

On support-staff access (older document)
“A restricted group of product support experts with appropriate accesses will be able to review customer usage data”

Lexis+ AI Security Information (Australian site), Q&A: Can LexisNexis employees see my chat queries?, Copyright 2023; file revised November 7, 2024. Read the source

Your data, in plain English

How long it is kept
  • LexisNexis says conversations are deleted automatically after 90 days.
  • This comes from its French-edition help page (July 29, 2026) and older U.S. and Australian documents.
  • The current U.S. product page gives no period for conversations.
  • Up to 10 documents uploaded without saving to a Vault are purged at the end of each session.
  • Vaults are kept until the user deletes them.
  • Results from Protégé Vault are kept 90 days in "My Conversations." After the subscription ends, LexisNexis may delete uploaded files once at least 60 days have passed.
  • The terms do not require it to.
  • Under the DPA, the customer can ask for personal data to be returned or deleted.
What happens after you delete
LexisNexis's help page (French edition) says a deleted conversation cannot be recovered. The General Terms say each user controls prompt history and its deletion. That control is subject to LexisNexis's automatic deletion controls, retention schedules, and applicable law. LexisNexis does not say how fast deleted data leaves its servers and backups. A delete-all option is not documented.
Copies kept anyway
  • The General Terms make a user's deletion subject to LexisNexis retention schedules and applicable law. LexisNexis does not publish those schedules.
  • LexisNexis says on its trust center that its model providers keep no prompts, under a zero-retention contract. This is not in the contract.
Temporary chat
Not offered as a named mode. Up to 10 documents uploaded without saving to a Vault are purged at the end of each session.
Memory
Protégé personalizes results to a user's preferences, jurisdiction, practice area, and document history. LexisNexis says users can modify or disable personalization at any time. It says customers decide whether personalization is enabled. LexisNexis does not say whether it is on by default.
Who at the provider can read it
  • The General Terms do not address LexisNexis staff access to prompts.
  • An older LexisNexis security document (2023, revised November 2024, on its Australian site) says a restricted group of product support experts can review customer usage data.
  • It says this is for product support and troubleshooting, and that the customer's identity is pseudonymized.
  • Under the DPA, people who process personal data must be bound by confidentiality.
  • LexisNexis says on its trust center that its model providers keep no prompts for review.
  • This is not in the contract.
Who else receives it
  • LexisNexis lists these subprocessors for file uploads in Lexis+ with Protégé (list updated October 20, 2026): Amazon Web Services, Anthropic (USA), Microsoft, Mistral AI (Sweden), OpenAI (USA), and Google.
  • Google's location is listed as "Global," and LexisNexis says it may vary.
  • The DPA says LexisNexis posts subprocessor changes 14 days in advance.
  • The customer may object.
  • The U.S. Privacy Laws Addendum forbids LexisNexis to sell personal information it collects under the agreement.
  • It also forbids LexisNexis to share that information.
  • No advertising use is stated.
  • If a law or court order requires disclosure, the General Terms require advance notice so the customer can seek a protective order.
  • Harvey customers who use Ask LexisNexis inside Harvey send that data to LexisNexis in the US.
Contract
  • LexisNexis General Terms and Conditions (effective October 19, 2026), the LexisNexis Data Processing Addendum (updated October 20, 2026) with its U.S. Privacy Laws Addendum (updated January 1, 2026), and the customer's order form.
  • The no-training promise binds LexisNexis.
  • It covers training large language models.
  • The DPA makes LexisNexis a processor for personal data it handles for the customer.
  • LexisNexis may change the General Terms by posting a new version.
All sources for this plan (17)

Last checked against LexisNexis's pages on October 25, 2026.

Not yet confirmed
  • The current retention period for U.S. conversations. The 90-day figure comes from LexisNexis's French-edition help page and older documents.
  • Whether the zero-retention contract with model providers, described on LexisNexis's trust center, covers every provider on the subprocessor list, including Google and Mistral.
  • How LexisNexis uses de-identified AI interactions to improve performance.
  • Whether personalization is on by default.
  • Whether the older document's statements on support-staff access and administrator controls are still current.
  • Whether firm administrators can set a retention period shorter than 90 days.
Protégé General AIYellowRated Yellow. The same General Terms apply as for Lexis+ with Protégé. LexisNexis promises not to use your prompts and files to train large language models and must treat them as confidential. A Data Processing Addendum makes LexisNexis your processor for personal data. What the plan lacks is organization control of retention: each user controls their own prompt history. Your prompts go to models from OpenAI, Google, or Anthropic. LexisNexis says on its trust center that these model companies are under a zero-retention contract. That promise is not in the contract.

Protégé General AI (inside Lexis+ with Protégé). Lexis+ with Protégé users who switch from Protégé Legal AI to Protégé General AI. General AI runs general-purpose models from OpenAI, Google, and Anthropic. Price: Part of a Lexis+ with Protégé subscription (custom quote).

Default settings Yellow

The same General Terms apply as for Lexis+ with Protégé, including the large-language-model no-training promise and confidential treatment. Each user controls their own prompt history; the firm has no published retention setting. Your prompts go to a model from OpenAI, Google, or Anthropic, or to a Best Fit mode that picks one. LexisNexis says on its trust center that its model providers are under a zero-retention contract. This is not in the contract.

Settings that lower the risk

  1. Protégé General AI on or offWhat it does: LexisNexis says users can switch General AI on or off. It says firms can turn it off for all users.
  2. Grounding sources (LexisNexis, web, and customer content)What it does: LexisNexis says users can ground answers in any combination of LexisNexis, web, and customer content. Deep Research works only when web sources are selected.
  3. Opt-out of AI failover (General Terms §4.6)What it does: If LexisNexis's cloud provider fails, LexisNexis may move AI processing to a backup location, which may be in another country. The firm can opt out by giving 30 days' written notice. A firm that opts out waives claims for the resulting loss of service.

Worth knowing

  • The Web Search Agent brings open-web results into answers. LexisNexis does not name the web search provider it uses.
  • On October 25, 2026, LexisNexis listed these General AI models: GPT-4o, OpenAI o3, GPT-5, GPT-5.1, Claude Sonnet 4, Claude Sonnet 4.5, and Gemini 3 Pro. It says new models are added over time.
  • Anthropic keeps data 30 days for its newest Fable 5 models unless zero retention applies. No Fable model was listed in General AI on October 25, 2026.
  • Unless the firm opts out by 30 days' written notice, AI processing may fail over to a backup location in another country.

What the policy says

On which model companies receive prompts
“Protégé General AI includes access to large language models from OpenAI, Google, and Anthropic”

General AI in Lexis+ with Protégé, FAQ: What models are available in Protégé General AI?, Accessed October 25, 2026. Read the source

On no training (binding)
“will not be used by LN to train any large language model(s)”

LexisNexis General Terms and Conditions, §4.3(b), Effective October 19, 2026. Read the source

On promises apply to web-sourced content (binding)
“apply regardless of whether a particular AI Service retrieves or surfaces web-sourced content”

LexisNexis General Terms and Conditions, §4.3(b); §4.3(c), Effective October 19, 2026. Read the source

On model-provider zero retention (trust center, not contract)
“we have a zero-retention contract with our foundation providers, ensuring they do not retain any prompts or other data for review and analysis.”

LexisNexis Trust Center, Knowledge Base (FAQ): LLM providers and human reviewers, Accessed October 25, 2026. Read the source

On firm can turn the mode off
“firms can turn off General AI for users entirely to meet firm or regulatory policies”

LexisNexis Announces Commercial Availability of Protégé General AI in Lexis+ AI (press release), User and Org-level Control, October 17, 2025. Read the source

On Web Search Agent; Best Fit; source selection
“the Web Search Agent that brings relevant open web insights into Protégé to expand responses”

LexisNexis Unveils Next-Generation Protégé General AI (press release), Agentic infrastructure; Flexible source selection, December 10, 2025. Read the source

On Anthropic 30-day retention for Fable 5 models
“we introduced 30-day data retention starting with Fable 5”

Developing Enterprise Frontier Safeguards with our customers (Anthropic), Solving the dilemma of frontier security, September 1, 2026. Read the source

On processor role
“process the personal data only on your documented instructions”

LexisNexis Data Processing Addendum, §2.3 (incorporated by General Terms §10.3), Last updated October 20, 2026. Read the source

Your data, in plain English

How long it is kept
  • LexisNexis says conversations are deleted automatically after 90 days.
  • This comes from its French-edition help page (July 29, 2026) and older U.S. and Australian documents.
  • The current U.S. product page gives no period for conversations.
  • Up to 10 documents uploaded without saving to a Vault are purged at the end of each session.
  • Vaults are kept until the user deletes them.
  • Results from Protégé Vault are kept 90 days in "My Conversations." After the subscription ends, LexisNexis may delete uploaded files once at least 60 days have passed.
  • The terms do not require it to.
  • Under the DPA, the customer can ask for personal data to be returned or deleted.
  • LexisNexis says on its trust center that its model providers keep no prompts.
  • This is not in the contract.
What happens after you delete
LexisNexis's help page (French edition) says a deleted conversation cannot be recovered. The General Terms say each user controls prompt history and its deletion. That control is subject to LexisNexis's automatic deletion controls, retention schedules, and applicable law. LexisNexis does not say how fast deleted data leaves its servers and backups. A delete-all option is not documented.
Copies kept anyway
  • The General Terms make a user's deletion subject to LexisNexis retention schedules and applicable law. LexisNexis does not publish those schedules.
  • LexisNexis says on its trust center that its model providers keep no prompts, under a zero-retention contract. This is not in the contract.
Temporary chat
Not offered as a named mode. Up to 10 documents uploaded without saving to a Vault are purged at the end of each session.
Memory
Protégé personalizes results to a user's preferences, jurisdiction, practice area, and document history. LexisNexis says users can modify or disable personalization at any time. It says customers decide whether personalization is enabled. LexisNexis does not say whether it is on by default.
Who at the provider can read it
  • The General Terms do not address LexisNexis staff access to prompts.
  • An older LexisNexis security document (2023, revised November 2024, on its Australian site) says a restricted group of product support experts can review customer usage data.
  • It says this is for product support and troubleshooting, and that the customer's identity is pseudonymized.
  • Under the DPA, people who process personal data must be bound by confidentiality.
  • LexisNexis says on its trust center that its model providers keep no prompts for review.
  • This is not in the contract.
Who else receives it
  • Prompts go to models from OpenAI, Google, and Anthropic.
  • A Best Fit mode picks the model by default.
  • The Web Search Agent draws on the open web.
  • LexisNexis does not name its web search provider.
  • The same subprocessors, DPA terms, and court-order notice apply as for Lexis+ with Protégé.
  • The U.S. Privacy Laws Addendum forbids LexisNexis to sell personal information it collects under the agreement.
  • No advertising use is stated.
Contract
The same General Terms (effective October 19, 2026) and Data Processing Addendum as Lexis+ with Protégé. The no-training promise is binding. It expressly applies when an AI service draws on web content. The contracts between LexisNexis and its model providers are not published. LexisNexis says on its trust center that they require zero retention.
All sources for this plan (11)

Last checked against LexisNexis's pages on October 25, 2026.

Not yet confirmed
  • Whether General AI costs extra beyond the Lexis+ with Protégé subscription.
  • Whether the trust center's zero-retention statement covers each General AI model, including Google's Gemini 3 Pro.
  • Which web search provider the Web Search Agent uses, and on what terms.
  • On what terms LexisNexis would offer any Anthropic model subject to 30-day retention, such as the Fable 5 family.

Our criteria

How we rate

Each color answers one question: may a D.C. lawyer put client confidences and secrets into this plan? Each plan has one color, set by its contract. The settings in each entry lower the risk but do not change the color. The full method, with the ethics rules behind each test, is on the tracker page.

Green

Fit for client confidences and secrets, after your usual vendor checks. All five are true:

  1. (1)A contract forbids the provider to train AI models on your content or use it for its own purposes.
  2. (2)A contract makes your content confidential and requires notice of a subpoena or government demand.
  3. (3)Staff and subcontractors see it only to run, support, or secure the service. Any outside AI model company is bound by the same terms.
  4. (4)Your organization can delete it or set how long it is kept.
  5. (5)The contract requires notice of a security breach, and an independent audit (SOC 2 Type II or ISO 27001) covers the service.
Yellow

Use for client confidences only after you deal with the gap. A business contract forbids training and limits use, but at least one Green condition is missing. Common gaps:

  1. (1)Your organization cannot delete content or set retention.
  2. (2)An outside AI model company works under its own terms.
  3. (3)The contract lets the provider use your content, even de-identified, for its own purposes.
  4. (4)The contract does not treat your content as confidential.
  5. (5)The contract bars personal data, which client files contain.
Red

Keep client confidences and secrets out. At least one of these is true:

  1. (1)The provider trains AI models on your content, or its terms let it.
  2. (2)Your content is used to choose ads.
  3. (3)Only consumer terms apply, so no contract keeps your content confidential.

Turning training off does not change the color. The only exception is the client's informed consent (D.C. Rule 1.6(e)(1)), and D.C. Bar Opinion 388 says clients typically should not be asked to consent to disclosures this broad.

Last checked against the provider's own pages on October 25, 2026. Quotations are kept short and link to the full page. Where the provider does not say something, we say so rather than guess. Ratings describe each provider's published standard terms; large customers often sign negotiated agreements that may differ. General information prepared by the BADC Young Lawyers Section AI Committee, not legal advice. Spotted an error or a change? Write to YLS@badc.org with the subject "Tracker correction."

Questions about the AI Committee, or want to get involved? Write to us directly.

YLS@badc.org