Does Fastmail train on your data?
Fastmail's privacy policy now says it in one sentence: it does not and has never trained AI on customer data. It was added on 2 September 2026, a week after Fastmail told us so by email.
No opt-out found in this policy
"Fastmail does not and has never trained AI on customer data."
- 2 Sept 2026
- 5 Oct 2026
- All plans
- global
- Fetched from the company
- 7C2B8085…185A
How we read it: Regraded UNCLEAR to B on 2 September 2026, the day the sentence appeared — the first policy any company has changed after a letter from this site. Fastmail said it by email on 27 August; the entry stayed UNCLEAR because a customer could not cite an email, and the follow-up asked for the sentence to go in the policy. It went in under a new heading, "3rd Party AI connections", beside the older line that the Anthropic connection runs "with model training switched off". B rather than A because it is a commitment rather than a stated inability, and because it sits under a third-party heading a reader has to look past.
train, training, opt out, switched off, AI
What the document does offer
- The new section, "3rd Party AI connections", captured 2026-09-02: the clause quoted here, and beside it "Fastmail only allows connections via the Anthropic API with model training switched off."
- The same section says customers "may choose, however are not required" to enter personal information into a connected third-party AI process — the connection is the customer's choice
- Before 2 September the policy carried only the Anthropic sentence, which is about the support assistant; the 17 August capture is kept and the change is on /changes
Why none of it applies: Nothing is on by default that an opt-out would need to reach. The commitment is general — "customer data", not one product or one path — and it is now in the document a customer would cite.
What Fastmail said when we showed them this
Changed“I have actually now added this line into our privacy policy. You can see this in the section "3rd Party AI connections".”
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.
- Nothing to switch off
- Not stated
- 18 / 100
- Not stated
- Unknown
How to opt out of Fastmail AI training
Plan by plan, from what Fastmail'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.
All plans
Nothing to switch off: Fastmail's policy says training is off here.
What this policy has done since we started watching
1 change detected since 17 Aug 2026. Each one is a diff between two archived copies, and each links to the passage that moved.
Everything we read
Vendor response
Their reply is quoted in full at the top of this page.
We wrote back on 28 Aug 2026, and they answered.
source ↗Show this verdict on your site
Fastmail is graded B, and a company that's earned that can say so with the verdict itself. The image below links back to this page and shows the live grade. If the grade changes, the badge changes the same day, wherever it's embedded. It's free and it isn't for sale.
