Companies Are Throttling Employees' AI Use Because It's Too Expensive

- AI tools like ChatGPT, GitHub Copilot, and other generative AI services charge companies per-seat (monthly per user) or per-token (per query/response). When thousands of employees use these tools heavily, costs can skyrocket — far beyond the initial per-user subscription fees, especially for API-based usage. - "Throttling" means deliberately limiting or capping how much an employee can use these AI tools (e.g., restricting the number of queries per day, blocking certain features, or placing users on a slower tier of the service). - This marks a shift from the 2023–2024 "AI gold rush," where companies encouraged widespread experimentation, to a phase of cost containment — similar to how firms once capped cloud computing or Slack storage usage after initial enthusiasm. - The article likely discusses enterprise AI governance tools and policies (e.g., Microsoft Copilot usage limits, enterprise ChatGPT admin controls) that employers are now deploying to prevent budget overruns from unchecked AI consumption.