Does Perplexity train on your data?
Perplexity trains on your searches by default on Free, Pro and Max. The switch is real and free — but it is in a help article, not the privacy policy, and it does not reach data already taken.
A self-serve opt-out exists
"Improve or create services and products, including our AI models."
- 7 Sept 2026
- 5 Oct 2026
- Free & Pro (queries, prompts and uploads)
- global
- Fetched from the company
- 949C31ED…A188
How we read it: Corrected 20 Aug 2026 after Perplexity replied. This page said declining meant invoking the objection right; that was wrong. A help-centre article names a toggle — Account settings, Preferences, "AI data retention" — and states it is "enabled by default" for Free, Pro and Max. Still D rather than C on the rule Figma is held to: the switch is real and free, but the privacy policy names neither it nor the default, so a reader of the policy alone would conclude there is no way out. The article also answers the question most companies leave open, and answers it against you: "Previously collected training data cannot be deleted or removed".
opt out, opt-out, object to, objection, legitimate interest, restrict, settings, choices
What the document does offer
- help centre, "Data Collection at Perplexity" (16 Jul 2026): "Some data may be used to train AI models and improve search quality, unless you opt out through Account Settings"
- same article: "For Free, Perplexity Pro and Perplexity Max users, AI Data Retention is enabled by default", with the route — Account settings, Preferences page, "AI data retention" toggle
- same article: "Opt-outs only apply to data collected after the opt-out date" and "Previously collected training data cannot be deleted or removed"
- in the privacy policy: "Objection to or Restriction of Certain Processing. In certain circumstances, you may object to the processing of your personal data" — no mention of the toggle anywhere
Why none of it applies: The operative control is the AI data retention toggle, published only in the help centre; the objection right this entry previously rested on is still in the policy, still reads "in certain circumstances", and is not the route a user would take. Perplexity's first reply called it an "AI Data Usage toggle"; asked about the mismatch, it accepted the help- centre wording as "the more precise wording for the control". Enterprise is excluded from training outright.
"We do not use the content of emails to create, train, improve, or fine-tune AI models."
- 7 Sept 2026
- 5 Oct 2026
- Email Service Information
- global
- Fetched from the company
- 949C31ED…A188
How we read it: Kept as its own tier because it is a real exclusion and the only one Perplexity writes out, but note how narrow it is: it covers the content of emails and nothing else. Its emphatic framing cuts both ways — a sentence drafted as an exception to the rest of the policy invites the reading that the rest of the policy does the opposite, which is what the tier above records.
What Perplexity said when we showed them this
Clarified“So yes: my earlier phrasing of "AI Data Usage toggle" was not the best label for the published control. The current help-center wording uses "AI Data Retention" [...] On the retention point, your reading matches the published guidance. I do not have a separate route to name for removing training data that was already collected before the opt-out. The available control applies going forward.”
What each plan can refuse
The same questions, asked of every plan we could identify. Every cell comes from a quoted clause or a documented search - see the evidence above.
- Self-serve toggle
- Not stated
- 53 / 100
- Not stated
- Unknown
- Nothing to switch off
- Not stated
- 18 / 100
- Not stated
- Unknown
How to opt out of Perplexity AI training
Plan by plan, from what Perplexity's own policy says about refusing. Where it publishes the steps, they're listed here with the limits they carry. Where there's nothing to switch off, or nothing graded, this says so instead of inventing a menu path.
Free & Pro (queries, prompts and uploads)
Yes. Perplexity has a self-serve setting.
- Open your Account settings.
- Go to the Preferences page.
- Switch off the AI data retention toggle.
Applies to data collected after you switch it off. Perplexity says earlier training data can't be removed.
Email Service Information
Nothing to switch off: Perplexity's policy says training is off here.
What this policy has done since we started watching
2 changes detected since 2 Jul 2026. Each one is a diff between two archived copies, and each links to the passage that moved.