# WeTransfer removed the machine-learning clause from its Terms of Service after user backlash (July 2025)

> After users read a newly added clause 6.3 as permission to train AI on their uploaded files, WeTransfer rolled the change back on 15 July 2025 — removing the machine-learning language and stating it does not use AI to process content shared via the service. Documented first-party by WeTransfer's own blog post and independently by The Register and Malwarebytes Labs, all three captures byte-tied to Wayback.

- Verification status: verified
- Case type: deployment
- Provider: WeTransfer (Bending Spoons) (https://wetransfer.com/)
- Client: WeTransfer, File-sharing / consumer SaaS (named)
- Sector: saas / NL / ops
- Verified on: 2026-08-24
- Canonical URL: https://theinternetninja.com/stories/wetransfer-ai-training-clause-terms-removed-after-backlash-2025/
- Source: The Internet Ninja (theinternetninja.com), independent verified-proof platform

## Verification method

WeTransfer's own walk-back blog post (Tier-1 first-party, byte-tied to its Wayback capture) plus two mutually independent newsrooms — The Register and Malwarebytes Labs — that carry the rollback and the company's statement firsthand and are each byte-tied to their Wayback snapshots (sha1-b32 == CDX digest). Every quoted fragment grep-F byte-verified against the decompressed captures.

## Full case file

## The problem

WeTransfer, the consumer file-sharing service, quietly updated its Terms of Service in early July 2025. Sometime after 2 July it rewrote clause 6.3 — Malwarebytes Labs reported that "Sometime after July 2, it updated clause 6.3 of the document" ([source](https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)) [Source 3]. The added language granted WeTransfer a "perpetual, worldwide, non-exclusive, royalty-free, transferable, sub-licensable license to use" ([source](https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)) user Content [Source 3]. In mid-July, creatives and other users read the clause as permission to train AI on the files they upload, and the criticism spread across social media.

## What was built

This is a Terms-of-Service reversal, not a deployment. In its own blog post WeTransfer quoted the disputed licence, which extended to "improve performance of machine learning models that enhance our content moderation process" ([source](https://web.archive.org/web/20250718202036id_/https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025)) — the same clause language The Register also reported [Source 1, Source 2]. The revised terms were set to take effect on 8 August 2025 (a date carried by the press, not the primary — see the verification box) [Source 2].

## The outcome

- **The clause was rolled back.** After the backlash WeTransfer reversed the change. In its first-party blog post the company said it had "removed the mention of machine learning" ([source](https://web.archive.org/web/20250718202036id_/https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025)) [Source 1]. The Register reported the company "rolled back changes it had introduced to its Terms of Service after they deeply upset users" ([source](https://www.theregister.com/2025/07/18/llm_products_terms_of_service/)) [Source 2], and Malwarebytes reported that WeTransfer had "rolled back language that allowed it to train machine learning models on any files that its users uploaded" ([source](https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)) [Source 3].
- **WeTransfer's framing.** In the same blog post WeTransfer stated firsthand that it does not use "machine learning or any form of AI to process content shared via WeTransfer" ([source](https://web.archive.org/web/20250718202036id_/https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025)) and added: "We apologize for any confusion this may have caused" ([source](https://web.archive.org/web/20250718202036id_/https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025)) [Source 1]. The Register separately quoted the company saying it does not use "any form of AI to process content shared via WeTransfer" ([source](https://www.theregister.com/2025/07/18/llm_products_terms_of_service/)) [Source 2]; per Malwarebytes the company said the clause "may have caused confusion for our customers" ([source](https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)) and that it does not "sell content to third parties" ([source](https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)) [Source 3].

**The weakest load-bearing source, named where you meet it:** WeTransfer's own blog post (Source 1) is a Tier-1 first-party primary, but it is the *subject* narrating its own reversal — an admission-against-interest (it removed the clause and apologised), not an independent audit. It does not stand alone: the rollback itself is independently reported by The Register (Source 2, Tier 2) and Malwarebytes Labs (Source 3, Tier 2), the two independent secondaries around the first-party origin. WeTransfer's counter-claim — that the machine-learning reference was only ever intended for possible future AI-powered content moderation, not a built or used feature — is recorded here as the company's position, not asserted as fact.

Why this matters for TIN: an AI-training permission read into a licence clause, then withdrawn under pressure, is exactly the kind of dated, publicly refutable reversal the verified-story model is built to pin down — the record supports the walk-back; nothing here depends on WeTransfer confirming it privately.

## How this was verified

- **Method.** Every quoted line was read verbatim from the captures in `sources/`; no figure is paraphrased or recalled. The WeTransfer blog is client-rendered (Prismic), so its article body lives in an embedded JSON payload — the four first-party fragments were bound against both the decompressed capture and a `.nextdata.txt` extract of that payload (each `grep -F` count ≥ 1).
- **Primary artifact.** WeTransfer's walk-back blog post (capture `20250718202036`, Tier-1 first-party), fetched via the Wayback `id_` raw payload, is byte-tied: sha1-base32 of the raw gzip bytes `PKMC2GUZKIAZHBVP42ZBPOSMACOXQKAB` == the live CDX digest at that snapshot. `first_publication_date` reads `2025-07-15T15:34:04+0000` inside the capture.
- **Independence.** Two mutually independent newsrooms carry the rollback and WeTransfer's statement firsthand — The Register (capture `20250718141105`, digest `GYKMWMV3VVFYJHSQ7VXFDTY64HNW2ESJ`) and Malwarebytes Labs (Danny Bradbury, capture `20250717143258`, digest `34YOTJJRSYTYP3GD2CPRFKDAXU4D6XOC`), each byte-tied to its snapshot. Aggregator reprints (AOL/Yahoo, Business Standard, PPC.land, and others) recycle the same text and are not counted as independent origins.
- **Conflation guards.** This is distinct from WeTransfer's separate 2024 acquisition by Bending Spoons and the associated layoffs, and from the other 2023–2025 ToS/AI-training reversals (Zoom, Dropbox, Adobe, Slack, LinkedIn) — none of which are cited here.
- **Open item (human, post-graduation).** The 8 August 2025 effective date and a canonical, archivable first-party copy of WeTransfer's statement are the human confirmation steps in `dossier.json → path_to_green`; the badge is amber `pending` and green never depends on WeTransfer confirming anything.
- **Verified on:** 2026-08-08 (presentation rebuilt 2026-08-19).

## Related case files

- [Zoom reversed its Terms-of-Service AI-training clause after backlash](/stories/zoom-terms-of-service-ai-training-customer-content-reversed-after-backlash-2023/) — the same shape two years earlier: a licence-clause rewrite read as consent to train AI on customer content, withdrawn under public pressure.
- [Dropbox walked back its third-party-AI (OpenAI) toggle](/stories/dropbox-third-party-ai-openai-toggle-backlash-data-not-used-to-train-models-2023/) — another file-storage service forced to clarify, publicly, that user files were not being used to train models.
- [Adobe added an AI-training clarification to its Terms of Use after backlash](/stories/adobe-terms-of-use-ai-training-clarification-added-after-backlash-2024/) — a creative-tools platform making the opposite move, adding explicit language to quell the same fear the WeTransfer clause raised.
- [Slack updated its privacy principles after an ML-training opt-out backlash](/stories/slack-ai-privacy-principles-updated-after-user-backlash-ml-training-opt-out-2024/) — a SaaS sibling in the same vein, where the dispute turned on default machine-learning use of customer data.

## Sources

1. **[Tier 1]** WeTransfer · "Terms of Service changes: July 2025" (first-party blog post; `first_publication_date` 2025-07-15) · 15 July 2025 · https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025 (archived: https://web.archive.org/web/20250718202036id_/https://wetransfer.com/blog/story/wetransfer-terms-of-service-changes-july-2025)
2. **[Tier 2]** The Register · "WeTransfer backtracks on TOS changes that opened the door to LLM training" · 18 July 2025 · https://www.theregister.com/2025/07/18/llm_products_terms_of_service/ (archived: https://web.archive.org/web/20250718141105/https://www.theregister.com/2025/07/18/llm_products_terms_of_service/)
3. **[Tier 2]** Malwarebytes Labs (Danny Bradbury) · "WeTransfer walks back clause that said it would train AI on your files" · 17 July 2025 · https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files (archived: https://web.archive.org/web/20250717143258/https://www.malwarebytes.com/blog/news/2025/07/wetransfer-walks-back-clause-that-said-it-would-train-ai-on-your-files)