Does Poe train on your data?
Poe's privacy policy says developers of third-party bots and apps "may view and store your chats on their servers to train their models". For Poe's own bots it says nothing either way.
No opt-out found in this policy
"Developers that create bots and apps on Poe using APIs may view and store your chats on their servers to train their models."
- 6 Sept 2026
- Chats with third-party developer bots and apps
- global
- Saved by hand — not re-read automatically
- A85DC757…9A47
How we read it: Re-graded from UNCLEAR on a sentence the first reading missed: the count note had attributed the one "train" in the policy to a user-research paragraph, and it is this one. The clause is about third-party developers, who may keep and train on the chats you have with their bots; the policy names no way to know which bots those are or to keep a chat out, and it does not say what Poe's own bots do. Poe's privacy centre, which explains a shield icon marking bots that train, sits on a path robots.txt disallows to this monitor and is not read here; medium confidence for that reason. Poe was written to on 2026-09-11 about the silence, before this sentence was found.
opt out, opt-out, object, settings, shield, train, default
What the document does offer
- in the policy: "Developers that create bots and apps on Poe using APIs may view and store your chats on their servers to train their models." — the clause; no control beside it
- in the policy: providers and developers may receive your interactions "to provide and generally improve their services, which they may process in their legitimate business interests, or based on the settings in your Poe account" — settings named for the providers' use, not for developers' training
- in the policy: "Please exercise caution and avoid sharing sensitive personal information with the bots and apps" — advice, not a control
- "opt out", "opt-out", "shield", "default": 0 matches in the policy
Why none of it applies: No route is stated in the policy. The only choices it names are account-level and for the model providers' legitimate interests; nothing marks which bots are third-party developers' or lets a chat with one be kept from training. The privacy centre that does mark them is disallowed to the monitor and uncaptured, so it cannot carry the grade. Scored as no opt-out on the document that can.
The policy's training sentence is about third-party developers. For the bots Poe itself runs on the providers' models, it says only that the providers may receive your chats "to provide and generally improve their services", which is not a training statement. The privacy centre that answers this for official bots is disallowed to the monitor and unread here.
- Poe Privacy Policy (poe.com/privacy), saved by hand 2026-09-06, 18 paragraphs
- train, training, fine-tune, machine learning, artificial intelligence, AI model, large language model
- 1 of 1 raw hits
- 14 Sept 2026
The one "train" is the developer sentence quoted on the other tier; it is about third-party developers' models, not Poe's own bots.
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.
- None offered
- Not stated
- 88 / 100
- Not stated
- Unknown
- Not established
- Not stated
- —
- —
- Not established
How to opt out of Poe AI training
Plan by plan, from what Poe'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.
Chats with third-party developer bots and apps
No opt-out is stated in the policy on file for Poe.
Poe's own bots
Not graded, so no route is pointed at. The policy's training sentence is about third-party developers. For the bots Poe itself runs on the providers' models, it says only that the providers may receive your chats "to provide and generally improve their services", which is not a training statement. The privacy centre that answers this for official bots is disallowed to the monitor and unread here.
What this policy has done since we started watching
The record starts here. Poe (Quora)'s policy was first captured on 6 Sept 2026, and there's no earlier copy to compare it against, so this verdict has no history behind it yet. From now on every re-read is diffed against that copy, and anything that moves appears here.
Everything we read
Vendor response
We wrote to Poe (Quora) on 15 Sept 2026, quoting the clauses above and inviting a correction. Nothing has come back. That's published because a company declining to comment on its own policy is something a reader should know, and because the invitation should be on the record whether or not it's taken up. Any reply goes here, word for word.
Report a wrong clause