AI labs are aggressively purchasing internal business data, including Slack messages and documents, to train autonomous AI agents, creating a booming but controversial market for workplace information.
Key Points
- Annual spending on data for AI training is projected to grow from under $1 billion in 2023 to nearly $6.5 billion by 2027.
- Startups like Micro1 are offering businesses between $100,000 and $2 million to license anonymized internal records for model development.
- Tech giants including Meta and SpaceX have faced internal backlash for attempting to harvest employee keystrokes and communications for AI training.
- Data brokers are using high-pressure tactics, such as 48-hour "exploding offers," to acquire proprietary information from small and midsize companies.
- The lack of federal privacy laws has created an ethically gray area regarding whether employee consent is required for the sale of workplace data.