Handshake AI pays up to $30K for work documents – but who owns them?

“It's unclear whether they're their own models, or whether or not they're going to be transferring this data,” Ahlering said. He also flagged potential copyright issues: “Who owns these documents? What vetting is there in place to make sure there's not confidential info?”

A Handshake AI representative did not respond to a request for comment on how the company would verify ownership. The position description does not define what makes a document “accepted,” leaving some candidates uncertain whether their submissions will yield a payout.

Ahlering said most employers would be uncomfortable with employees asking to sign over document rights. “I think most employers would be uncomfortable with that,” he said.

For professionals tempted by the $30,000 cap, the safest conclusion is that work created for an employer is off limits. Material created entirely outside of any employment or consulting relationship, and free of confidential business data, is the only clear candidate. Even then, the acceptance criteria remain opaque.

Handshake AI’s move fits a broader pattern. AI training firms have been paying for diverse datasets – from folding laundry videos to home-cleaning footage – as they race to improve machine-learning models. The demand for proprietary, high-quality text is especially intense as companies seek to train models on material that is not already scraped from the public web.

The document-contributor role offers a concrete price: $6 per page. But the legal and practical barriers to cashing in are steep, and Handshake AI has not said how it will police them.