Data Brokers Race to Buy Corporate Files and Emails to Train AI Models
A quiet boom is underway in the market for private workplace data. AI companies such as Anthropic and OpenAI are reportedly paying media firms between $350,000 and $2 million for archives of old work records, including documents, emails, Slack messages, and call transcripts. Data acquisition startups claim labs are spending billions to acquire private data now that public sources are exhausted.
Companies like Handshake are aggressively courting individual professionals as well, offering a freelance writer $1,000 for the files behind a single story and purchasing graduate students' papers and coursework in bulk. Another startup called Sunset pitches itself as a service that helps wind down startups, scrubbing and anonymizing internal data before brokering deals with AI labs. Cold emails pitching these data sales are now bleeding into ordinary inboxes.
Industry practitioners argue this business could affect workers and society as much as the massive data centers dominating headlines. The rush to monetize professional work raises unresolved questions about consent, compensation, and privacy, as employees' day-to-day communications become raw material for training the next generation of AI models.