WeTransfer almost used your files to train AI—without telling you 🤖

Cliquez ici pour lire en français

It’s the kind of update no one reads—until it blows up. In late June 2025, users of WeTransfer — one of the world’s most popular file-sharing platforms — noticed something alarming buried in the latest Terms of Service: your files could be used to “improve machine learning systems.” Translation: your data, used to train artificial intelligence.

The discovery sparked an immediate backlash, especially among creatives, privacy advocates, and tech professionals. Under pressure, WeTransfer swiftly removed the clause—but the damage was done. So what actually happened? And what does it mean for the rest of us?

TOS quietly updated… but not quietly enough 🔍

It all started when eagle-eyed users — including Mastodon user @notdan — spotted changes in WeTransfer’s Terms of Service. The new language, quietly published and set to take effect in July, included a clause stating that user data could be used “to develop AI models.”

The archived version from June 28, 2025, specifically stated that WeTransfer reserved the right to use user content “to improve the effectiveness of automated systems, including via machine learning.”

The problem? Files uploaded to WeTransfer often contain sensitive, personal, or copyrighted material. The idea that this data could be fed into an AI system—without explicit consent—felt like a major breach of trust.

Internet erupts: instant backlash ⚡

The news spread like wildfire across Reddit, X (formerly Twitter), and LinkedIn. Users started asking the obvious: had their creative work, private documents, or client files already been used to train AI? Could it happen in the future?

The backlash was particularly strong among designers, developers, PR professionals, and creatives—many of whom rely on WeTransfer for daily file sharing with clients. Some users publicly announced they were abandoning the platform, calling for more privacy-conscious alternatives.

Bye forever, WeTransfer.

— Rami Ismail (رامي) (@ramiismail.com) 2025-07-14T23:57:25.275Z

WeTransfer responds fast: “We hear you” 🎙️

In the face of mounting criticism, WeTransfer issued a rapid response. A company spokesperson told several tech outlets:

“We do not use machine learning or any form of AI to process content shared via WeTransfer, nor do we sell content or data to third parties.”

They clarified that the AI-related language was originally meant to improve automated moderation tools—not to mine user files for generative AI training or commercial use.

Still, in an effort to “avoid confusion or legitimate concern,” the controversial clause was pulled from the TOS immediately.

A trust wake-up call: what now? 🤔

Even though WeTransfer walked it back, a deeper question remains: can we still trust platforms that quietly add clauses like this to their terms—without proper disclosure?

Many users are now turning to more transparent or open-source file-sharing tools. A few alternatives gaining traction:

• SwissTransfer 🇨🇭 – free, secure, no ads, no tracking.
• SendAnywhere – easy to use with end-to-end encryption.
• Tresorit Send – European hosting with full encryption.
• OnionShare – for the ultra-cautious; anonymous file transfers via Tor.

But this is bigger than just WeTransfer. The incident raises a broader issue in the age of AI: when we “temporarily share” our data, does it still belong to us—or are we feeding the machines of the future?

Transparency or turbulence: time to choose

To its credit, WeTransfer backpedaled quickly. But this episode makes one thing clear: users aren’t passive anymore. Reading the fine print is now an act of digital self-defense, and hitting “send” is no longer a mindless action.

More than anything, this story marks a turning point. Default data surrender is no longer acceptable.

In today’s AI-infused internet, transparency isn’t a bonus—it’s the bare minimum if platforms want to earn and keep our trust.

So what do you think—was their U-turn enough? Or are you switching to a new tool? 🚀 Drop your favorite alternatives in the comments and let’s compare notes.