Research · Dated terms extract
How frontier labs use what you paste
OpenAI, Anthropic, Google, and SpaceXAI publish different rules for training, retention, and human review depending on the product you are actually using. This note records what those pages said on 27 August 2026. It is not a model ranking and it is not legal advice.
Source review 27 August 2026
Written and edited by Tyler Martin. Terms and pricing retrieved from the labs’ own pages on 27 August 2026. Drafting assistance from Grok 4.6 (xAI).
How to read it.
A row is an extract from a named page on a named date. Feature, region, and account settings can differ from the row. Silence on a point is treated as worst-case. Check the specific product and agreement before you paste anything you would mind seeing again.
The pattern
Consumer chat products at these four labs, including expensive individual plans, typically train on inputs unless the user changes a setting. Business, API, and enterprise agreements more often say they will not train by default. Logs, reviewers, and feature exceptions still exist on those paths. The published price is a weak proxy for which agreement you are on.
That split is a race as well as a contract problem. Once models are close enough in quality, a lab that can train on more private work has an advantage the others feel pressure to match. Each lab’s published terms are a move in that game, not a settled ethics statement.
This table is a rubric for reading the notes below. It does not replace any lab’s current terms.
Price is a weak proxy
OpenAI’s individual pricing comparison on 27 August 2026 puts Free, Go, Plus, and Pro on the same training row: “Content is used to train our models: Opt-out available.” Help Center copy lists Pro at $100 (5× Plus usage) and $200 (20×). Paying $200 does not move a ChatGPT account onto the Business Terms. Those terms cover ChatGPT Enterprise, the APIs, and other business products.
Google’s Gemini Apps at $0, Google AI Plus at $4.99, Google AI Pro at $19.99, and Google AI Ultra at $99.99 or $199.99 all sit under the Gemini Apps Privacy Notice, which tells users not to enter confidential information a reviewer might see. Claude Pro at $20 and Claude Max from $100 remain consumer opt-out on claude.com/pricing. Claude Team at $20 a seat is the first listed Claude plan whose pricing page says “No model training on your content by default.”
ChatGPT Plus’s exact dollar price and SuperGrok’s dollar price were client-rendered on the live pricing pages and are omitted rather than guessed.
Training is not the only copy
A no-training clause does not settle whether inputs are stored, who can read them, or whether a feature makes another copy. OpenAI’s API has not trained on API data by default since 1 March 2023; abuse logs still exist for 30 days unless Zero Data Retention is approved, and ZDR still has holes (videos, Assistants, files, live web search, CSAM image review). Anthropic’s commercial terms say they may not train on Customer Content; Fable 5 and Mythos 5 require 30-day retention and are unavailable under ZDR. Google Cloud will not train on customer data without permission, then lists Grounding with Google Search as storage that cannot be turned off if you use that feature. SpaceXAI’s consumer terms let a logged-in user turn off “Improve the model” and still grant a perpetual license to use User Content to provide and improve the Service.
What to paste
The terms below are why the same prompt is fine in one account and reckless in another. Put material in a bucket before it enters a lab product.
Public
Repeat it freely
Published articles, public code, public patents, press, and questions that would not hurt you if repeated aloud. Consumer chat is usually an acceptable place for this.
Private but replaceable
Judgment call
Rough notes, generic drafts, things you could rebuild. Read the note for that account before you paste.
Protected
Controlled only
Unfiled inventions, trade secrets, proprietary source, customer data, privileged analysis, deal terms, non-public research, and the prompts that hold accumulated know-how. Use a local model, a commercial API or enterprise environment whose exact terms you have checked, a redacted non-enabling version of the question, or a human conversation under confidentiality terms.
A patent application may make an invention public later by design. A trade secret only survives while you take reasonable steps to keep it secret. Casual disclosure into a consumer chat is how both fail.
Before you paste: would you be okay if this exact text appeared in a future training set, an internal review queue, a log, or later in a lawsuit? If the answer is no, strip it, shrink it, or use a controlled tool.
Lab notes, 27 August 2026
Twenty-seven product rows across four labs. Leakage risk here is a coarse editorial label for unfiled work, not a score. High means consumer terms with training on by default or a toggle, and possible human review. Medium means a no-training default with logs and feature exceptions. Lower means an enterprise or zero-retention path that still has named holes.
OpenAI
ChatGPT Free · High leakage risk
ChatGPT Free
Published price: $0 / month
Training: Trains unless you opt out
Retention / review: Account history until you delete it. Consumer Terms of Use (effective 1 Jan 2026) say OpenAI may use Content to provide, maintain, develop, and improve the Services. Human review: possible.
Feature exceptions: Memory; file uploads; browse and third-party apps; custom GPTs
Jurisdiction: US · OpenAI OpCo, LLC
Quoted language: “If you do not want us to use your Content to train our models, you can opt out by following the instructions in this article.”
Sources: OpenAI Terms of Use · ChatGPT pricing (privacy row)
ChatGPT Go · High leakage risk
ChatGPT Go
Published price: Paid · may include ads
Training: Same consumer opt-out as Free
Retention / review: Same consumer Terms of Use as Free and Plus. Pricing page says this plan may include ads. Human review: possible.
Feature exceptions: Ads; Memory; uploads
Jurisdiction: US · OpenAI OpCo, LLC
Quoted language: “Content is used to train our models: Opt-out available (all individual plans on the pricing comparison, including Go).”
Sources: ChatGPT pricing · OpenAI Terms of Use
ChatGPT Plus · High leakage risk
ChatGPT Plus
Published price: Paid consumer
Training: Same consumer opt-out as Free
Retention / review: Same consumer Terms of Use. The individual pricing comparison lists Plus under the same 'opt-out available' training row as Free and Pro. Human review: possible.
Feature exceptions: Memory; projects; custom GPTs; apps connecting to internal tools
Jurisdiction: US · OpenAI OpCo, LLC
Quoted language: “Our Business Terms govern use of ChatGPT Enterprise, our APIs, and our other services for businesses and developers.”
Sources: OpenAI Terms of Use · ChatGPT pricing
ChatGPT Pro · High leakage risk
ChatGPT Pro
Published price: $100 or $200 / month
Training: Same consumer opt-out as Free
Retention / review: Same consumer Terms of Use. Help Center (updated 25 Aug 2026): Pro $100 is 5× Plus usage; Pro $200 is 20× Plus usage. Human review: possible.
Feature exceptions: Memory; Codex; deep research; record mode
Jurisdiction: US · OpenAI OpCo, LLC
Quoted language: “Content is used to train our models: Opt-out available (Free, Go, Plus, and Pro on the individual pricing comparison).”
Sources: About ChatGPT Pro tiers · ChatGPT pricing · OpenAI Terms of Use
ChatGPT Business · Medium leakage risk
ChatGPT Business
Published price: Workspace, self-serve
Training: No training by default
Retention / review: Workspace admins control retention. Deleted or unsaved conversations removed within 30 days unless law or harm prevention requires longer. Specialized contractors may review for abuse. Human review: specialized contractors for abuse review.
Feature exceptions: Public GPT sharing if admins enable it; feedback opt-in
Jurisdiction: US · business terms + DPA available
Quoted language: “By default, we do not use your business data for training our models.”
Sources: Enterprise privacy at OpenAI (8 Jan 2026)
ChatGPT Enterprise / Edu / Healthcare · Lower leakage risk
ChatGPT Enterprise / Edu / Healthcare
Published price: Contract
Training: No training by default
Retention / review: Workspace admins control retention. Deleted conversations removed within 30 days unless legally required. Authorized employees access chats for incidents, recovery with permission, or law. Human review: limited to incidents, recovery with permission, or law.
Feature exceptions: Compliance API / audit logs; apps and connectors
Jurisdiction: US · SOC 2 Type 2; Healthcare BAA path
Quoted language: “We do not train our models on your data by default.”
Sources: Enterprise privacy at OpenAI (8 Jan 2026)
API Platform (default) · Medium leakage risk
API Platform (default)
Published price: Usage
Training: No training since 1 Mar 2023
Retention / review: Abuse-monitoring logs up to 30 days by default. Assistants, threads, vector stores, files, conversations, and videos retain application state until deleted or as documented. Prompt cache may live on GPU machines up to 24 hours. Human review: specialized contractors for abuse review.
Feature exceptions: /v1/videos not ZDR-eligible; Assistants and files retain state; live web search is not HIPAA/BAA; image/file CSAM scan can retain even under ZDR
Jurisdiction: US; data residency is a paid project option with a 10% uplift on eligible models from 5 Mar 2026
Quoted language: “As of March 1, 2023, data sent to the OpenAI API is not used to train or improve OpenAI models (unless you explicitly opt in).”
Sources: Your data (API docs) · Enterprise privacy at OpenAI
API Zero Data Retention · Lower leakage risk
API Zero Data Retention
Published price: Approval required
Training: No training; customer content excluded from abuse logs on eligible endpoints
Retention / review: Eligible endpoints skip abuse-log retention. Ineligible endpoints and capabilities still store application state. Eyes Off / Safety Retention can put some models back into logs. Human review: law or safety holds.
Feature exceptions: CSAM image/file review; ineligible endpoints (Assistants, files, videos, conversations); MCP servers are third-party retention
Jurisdiction: US; ZDR is per organization/project after approval
Quoted language: “Zero Data Retention excludes customer content from abuse monitoring logs in the same way as Modified Abuse Monitoring.”
Sources: Your data (API docs)
Anthropic
Claude Free · High leakage risk
Claude Free
Published price: $0 / month
Training: Trains unless you opt out
Retention / review: Deleted chats leave history immediately and backend storage within 30 days. If model improvement is on, de-identified copies may sit in training pipelines up to 5 years. Flagged chats: inputs/outputs up to 2 years; trust-and-safety scores up to 7 years. Feedback: 5 years. Human review: safety review and feedback.
Feature exceptions: Feedback (thumbs) still trains; safety-flagged chats still used; Incognito chats are excluded from model improvement; connectors not used unless pasted
Jurisdiction: US · Anthropic, PBC (consumer terms)
Quoted language: “We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings.”
Sources: Consumer Terms of Service · Claude pricing (model training row: Opt-out) · How long do you store my data? (1 Jul 2026)
Claude Pro · High leakage risk
Claude Pro
Published price: $20 / month, or $17 / month billed $200 annually
Training: Same consumer opt-out as Free
Retention / review: Same consumer retention as Free. Pricing page lists Free, Pro, Max 5x, and Max 20x all as 'Model training: Opt-out'. Human review: safety review and feedback.
Feature exceptions: Claude Code on consumer plans is still consumer terms; Cowork, Chrome, memory, connectors
Jurisdiction: US · Anthropic, PBC
Quoted language: “Even if you opt out, we will use Materials for model training when: (1) you provide Feedback … or (2) your Materials are flagged for safety review.”
Sources: Claude pricing · Consumer Terms of Service · Is my data used for model training? (consumer, 16 Mar 2026)
Claude Max · High leakage risk
Claude Max
Published price: From $100 / month
Training: Same consumer opt-out as Free
Retention / review: Same consumer terms as Free and Pro. Higher usage is not a different privacy contract. Human review: safety review and feedback.
Feature exceptions: Claude Code on Max is still a consumer plan; not ZDR-eligible
Jurisdiction: US · Anthropic, PBC
Quoted language: “Claude consumer products: Claude Free, Pro, and Max plans, including when customers on those plans use Claude's web, desktop, or mobile apps or Claude Code, are not ZDR-eligible.”
Sources: Claude pricing · API and data retention
Claude Team · Medium leakage risk
Claude Team
Published price: $20 / seat / month annual ($25 monthly); Premium $100 / $125
Training: No training by default
Retention / review: Commercial terms: Customer Content is Confidential Information; Anthropic may not train on it. Feedback still stored up to 5 years unless Rate chats is disabled. Team/Enterprise product interfaces are not ZDR-eligible. Human review: optional feedback.
Feature exceptions: Thumbs feedback; Claude Code through Team still follows commercial terms; product UI is not ZDR
Jurisdiction: US or Ireland entity depending on residence
Quoted language: “No model training on your content by default.”
Sources: Claude pricing (Team) · Commercial Terms of Service · Is my data used for model training? (commercial)
Claude Enterprise · Lower leakage risk
Claude Enterprise
Published price: $20 / seat + usage at API rates
Training: None by default
Retention / review: Custom data retention controls, audit logs, Compliance API, HIPAA-ready offering. Product interfaces still not ZDR-eligible except Claude Code through Enterprise with ZDR enabled. Human review: admin-controlled.
Feature exceptions: Compliance API / Activity Feed (Activity Feed retains 6 years); remote session transcripts 6 years; Covered Models require 30-day retention
Jurisdiction: US or Ireland; HIPAA BAA path
Quoted language: “Anthropic may not train models on Customer Content from Services.”
Sources: Claude pricing (Enterprise) · Commercial Terms of Service · API and data retention
Claude API (default) · Medium leakage risk
Claude API (default)
Published price: Fable 5 $10 / $50 per MTok; Opus 5 $5 / $25; Sonnet 5 $2 / $10; Haiku 4.5 $1 / $5. US-only inference 1.1×.
Training: No training without express permission
Retention / review: Conversation content is not retained by default except Covered Models (Fable 5, Mythos 5), which require 30-day retention. Batch jobs retain 29 days. Console playground is not ZDR. Human review: policy enforcement.
Feature exceptions: Covered Models require 30-day retention and are not ZDR; code execution, batch, managed agents, skills; third-party tools
Jurisdiction: US; US-only inference available at 1.1×
Quoted language: “Retained data is never used for model training without your express permission.”
Sources: API and data retention · Claude pricing (API) · Commercial Terms of Service
Claude API Zero Data Retention · Lower leakage risk
Claude API Zero Data Retention
Published price: Sales enablement, per organization
Training: No at-rest storage of prompts/responses after the response, on eligible features
Retention / review: Eligible /v1/messages features. Using a 'No' feature (batch, skills, managed agents, code execution) steps outside ZDR for that data; the API does not block it. Fable 5 and Mythos 5 are unavailable under ZDR unless a workspace turns on 30-day retention. Human review: flagged_and_legal_holds.
Feature exceptions: Covered Models; Claude Console; consumer products; Team/Enterprise UI; CORS blocked under ZDR
Jurisdiction: US; enablement does not auto-extend to other orgs on the same account
Quoted language: “Under a ZDR arrangement, Anthropic does not store customer prompts or responses at rest after the API response is returned.”
Sources: API and data retention
Google
Gemini Apps (Free) · High leakage risk
Gemini Apps (Free)
Published price: $0 / month with a Google Account
Training: Used to improve Google services, including generative models
Retention / review: Gemini Apps Activity auto-delete defaults to 18 months (3 / 36 / indefinite options). Human-reviewed chats are kept up to 3 years even after you delete activity. Keep Activity off still uses chats to respond and to protect Google, including with human reviewers. Human review: yes, including for model improvement.
Feature exceptions: Connected Apps and MCP tools; Gemini Spark remote browser cookies; Canvas apps (creator can see shared data); Gemini Live audio/video; signed-out use follows the Google Privacy Policy
Jurisdiction: Google LLC, or Google Ireland Limited in EEA/Switzerland
Quoted language: “Please don’t enter confidential information that you wouldn’t want a reviewer to see or Google to use to improve our services, including machine-learning technologies.”
Sources: Gemini Apps Privacy Hub (notice 29 Jun 2026; hub 10 Aug 2026) · Gemini subscriptions
Google AI Plus · High leakage risk
Google AI Plus
Published price: $4.99 / month
Training: Same Gemini Apps Privacy Notice as Free
Retention / review: Paid Google AI plan; subscription information is collected. The Gemini Apps Privacy Notice still governs chats. Human review: yes, including for model improvement.
Feature exceptions: Gemini in Chrome; Gmail/Workspace side panel; Flow credits
Jurisdiction: Google LLC / Google Ireland
Quoted language: “If you have a paid subscription to a Google AI plan, subscription related information [is collected].”
Sources: Gemini subscriptions · Gemini Apps Privacy Hub
Google AI Pro · High leakage risk
Google AI Pro
Published price: $19.99 / month
Training: Same Gemini Apps Privacy Notice as Free
Retention / review: Higher usage and 5 TB storage do not, on the public pages retrieved, replace the Gemini Apps Privacy Notice. Human review: yes, including for model improvement.
Feature exceptions: Jules; Antigravity; Gemini in Gmail/Docs; YouTube Premium Lite bundle
Jurisdiction: Google LLC / Google Ireland
Quoted language: “These uses extend to the generative AI models and other machine-learning technologies powering our services.”
Sources: Gemini subscriptions · Gemini Apps Privacy Hub
Google AI Ultra · High leakage risk
Google AI Ultra
Published price: $99.99 / month (5×) or $199.99 / month (20×)
Training: Same Gemini Apps Privacy Notice as Free
Retention / review: Highest consumer spend on the Gemini Apps path. Gemini Spark (remote browser and computer) is listed for Ultra in select countries and has its own data path. Human review: yes, including for model improvement.
Feature exceptions: Gemini Spark remote browser cookies and screenshots; Project Genie; YouTube Premium
Jurisdiction: Google LLC / Google Ireland
Quoted language: “Human reviewers (including trained reviewers from our service providers) review some of the data we collect for these purposes.”
Sources: Gemini subscriptions · Gemini Apps Privacy Hub
Gemini API / AI Studio (unpaid) · High leakage risk
Gemini API / AI Studio (unpaid)
Published price: Free quota / AI Studio
Training: Used to improve Google products, including ML
Retention / review: Human reviewers may read API input and output after it is disconnected from your account and API key. EEA/UK/Switzerland: paid-service data terms apply even to unpaid quota. Human review: yes, including for model improvement.
Feature exceptions: Grounding with Google Search extra logging; agentic Computer Use API; tuning data retained with tuned models
Jurisdiction: Google; EEA unpaid quota follows paid data terms
Quoted language: “Do not submit sensitive, confidential, or personal information to the Unpaid Services.”
Sources: Gemini API Additional Terms (effective 23 Mar 2026)
Gemini API (paid Cloud billing) · Medium leakage risk
Gemini API (paid Cloud billing)
Published price: Cloud Billing
Training: Does not use prompts or responses to improve products
Retention / review: Limited logging to detect Prohibited Use Policy violations. Processed under the Data Processing Addendum where Google is a processor. AI Studio counts as Paid if the account has an active Cloud Billing project or is a Workspace enterprise account. Human review: policy enforcement.
Feature exceptions: Grounding with Google Search still collects extra data; agentic services: you are responsible for actions
Jurisdiction: Google Cloud entity terms
Quoted language: “When you use Paid Services … Google doesn't use your prompts … or responses to improve our products.”
Sources: Gemini API Additional Terms
Gemini Enterprise Agent Platform (Google Cloud) · Lower leakage risk
Gemini Enterprise Agent Platform (Google Cloud)
Published price: Google Cloud
Training: Will not train or fine-tune on your data without prior permission
Retention / review: ZDR is a checklist, not a single switch. Abuse-monitoring prompt logs need an exception. Grounding with Google Search stores derived queries up to 3 days and cannot be disabled if you use it. Maps grounding stores 30 days. Interactions API store defaults to true. In-memory cache is 24 hours and is called compatible with ZDR; it can be disabled per project. Human review: abuse monitoring unless an exception is granted.
Feature exceptions: Grounding with Google Search; Grounding with Google Maps; Advanced AI Safety Addendum models; CodeMender session data up to 7 days; Gemini Live session resumption 24 hours if enabled
Jurisdiction: Google Cloud; CDPA
Quoted language: “Google won't use your data to train or fine-tune any AI/ML models without your prior permission or instruction.”
Sources: Gemini Enterprise Agent Platform and zero data retention (updated 27 Aug 2026)
SpaceXAI
Grok (logged in) · High leakage risk
Grok (logged in)
Published price: Free to try; SuperGrok is a paid upgrade (price is client-rendered)
Training: Trains unless you turn off Improve the model
Retention / review: History until you delete it. Deleted conversations and Private Chat are queued for deletion within 30 days unless law, safety, or de-identification applies. Authorized personnel may review conversations, including to improve model performance. Human review: yes, including for model improvement.
Feature exceptions: Feedback still used for training after opt-out; Grok on X is under X Terms, not these Terms; Agentic Actions can send content to third parties; share links are public
Jurisdiction: US · SpaceXAI LLC (Nevada); processing in the United States
Quoted language: “When logged into our Service, you can select whether or not you want us to use your User Content to improve our products and services and train our models.”
Sources: Terms of Service — Consumer (effective 24 Aug 2026) · Consumer FAQs · Privacy Policy (24 Aug 2026)
Grok (logged out) · High leakage risk
Grok (logged out)
Published price: $0
Training: No opt-out when unauthenticated (excluding EU/UK per FAQ)
Retention / review: FAQ: unauthenticated users will not have data retained for their later use after the session. Consumer Terms: where permitted, logged-out use grants full rights to use data for product development and model training. Human review: possible.
Feature exceptions: EU/UK excluded from this logged-out training rule per Consumer FAQs
Jurisdiction: US · SpaceXAI LLC
Quoted language: “If you do not log into your account to access Grok (i.e., you are unauthenticated), where permissible, we may collect and retain your content on an anonymous basis. As a result, in some regions (excluding the EU/UK), when you use Grok without logging in, you won’t have the option to opt out of model training.”
Sources: Consumer FAQs · Terms of Service — Consumer
Grok Private Chat · Medium leakage risk
Grok Private Chat
Published price: Setting on Free or paid
Training: Not used for model training
Retention / review: Not shown in history. Deleted from SpaceXAI systems within 30 days unless needed for legal, compliance, or safety. Human review: possible for safety, security, or legal holds.
Feature exceptions: Safety, security, and legal holds
Jurisdiction: US · SpaceXAI LLC
Quoted language: “When using Private Chat, where available, your content and interactions are not used for model training.”
Sources: Consumer FAQs · Privacy Policy
SpaceXAI API / Enterprise · Medium leakage risk
SpaceXAI API / Enterprise
Published price: Prepaid credits or invoice
Training: Will not train foundation models on User Content, subject to disclosures and customer-controlled settings
Retention / review: User Content deleted no later than 30 days after the session unless Order Form, docs, law, or safety requires otherwise. User Content is Customer Confidential Information. Human review: safety and compliance.
Feature exceptions: De-identified/aggregated data unless ZDR is on; Third-Party Services; Beta Offerings
Jurisdiction: Texas law; Wichita County, Texas courts
Quoted language: “SpaceXAI will not use any User Content to train any foundation models, large language models, or other artificial intelligence systems or to develop any new products, services, or features, subject to disclosures to Customer and Customer-controlled user settings.”
Sources: Terms of Service — Enterprise (14 Aug 2026)
SpaceXAI ZDR-Enabled API · Lower leakage risk
SpaceXAI ZDR-Enabled API
Published price: Enterprise election
Training: Transient inference only
Retention / review: User Content exists only transiently to generate the response; deleted on the earlier of one hour after inference or delivery of the response. No logs, backups, or durable copies. Customer is told SpaceXAI then cannot produce the data for subpoenas. Human review: none after deletion, by design.
Feature exceptions: Customer must send Personal Data only through the ZDR-Enabled API; PHI requires a BAA and ZDR; customer is solely responsible for copies
Jurisdiction: Texas; DPA at x.ai/legal/data-processing-addendum
Quoted language: “When and to the extent Customer elects to use ZDR, User Content will exist in SpaceXAI systems only transiently and solely to the extent required to generate and return the real-time response.”
Sources: Terms of Service — Enterprise (14 Aug 2026)
Method and limits
Pages were fetched on 27 August 2026 from the URLs listed in each note. Quotes are from those pages. ChatGPT Plus and SuperGrok sticker prices did not appear in the retrieved HTML (client-rendered) and are left blank rather than inferred. Perplexity, Mistral, DeepSeek, Microsoft Copilot, and Meta AI are not in this pass. Absence from the list is not a clean bill of health.
Terms move. OpenAI’s consumer Terms of Use are dated 1 January 2026. SpaceXAI’s consumer terms and privacy policy are dated 24 August 2026. Google’s Gemini Apps Privacy Hub was updated 10 August 2026. Anthropic’s consumer retention article is dated 1 July 2026. A later change on any of those pages is not reflected here until the note is edited.
Primary sources, 27 August 2026:
OpenAI Terms of Use;
ChatGPT pricing;
ChatGPT Pro tiers;
OpenAI enterprise privacy;
OpenAI API “Your data”;
Anthropic consumer terms;
Anthropic commercial terms;
Claude pricing;
Anthropic API retention / ZDR;
Anthropic consumer retention;
Anthropic consumer training;
Anthropic commercial training;
Gemini Apps Privacy Hub;
Gemini subscriptions;
Gemini API Additional Terms;
Google Cloud ZDR;
SpaceXAI consumer terms;
SpaceXAI consumer FAQs;
SpaceXAI privacy policy;
SpaceXAI enterprise terms.
This is an operational extract, not legal advice. Terms, configurations, and laws change. A note is a starting point, not a ruling.