Microsoft CEO Satya Nadella recently put out a blog post where he warns that companies using AI are basically paying twice. Once with actual money for token usage, and again by handing over data without fully realizing it. His point is that every prompt, every correction, every fix an employee makes while using these models becomes institutional knowledge that the model maker quietly absorbs.
src
What people dont really think about is how this can actually harm companies. Businesses are used to, or almost forced into, using models built by Anthropic or ChatGPT because that is simply where the tools are right now. And in doing so they may be feeding data that could later be used to build competitors against them or expose vulnerabilities they never intended to share. That is a real concern and Nadella is not wrong to raise it.
But to me, wasnt this the whole point of AI innovation in the first place. Creating smarter systems was always going to create competition somewhere down the line, that is how progress works. I think Nadella comes across as one sided here, mostly because Microsoft has its own interest in pushing companies toward Azure and its own ecosystem. AI isnt just about generating fake art or flashy demos, it has quietly become a real tool inside software houses, doing serious engineering work, and that side of the story gets ignored when the conversation turns into a warning about data ownership.