322
Audio & Video Production312
Software Development229
Automation & Workflow206
Writing & Content Creation188
Marketing & Growth175
AI Infrastructure & MLOps143
Design & Creative150
Photography & Imaging145
Data & Analytics110
Voice & Speech123
Education & Learning119
Sales & Outreach113
Customer Support111
Research & Analysis86
In a new book, Carissa Véliz argues generative AI mainly predicts patterns, and that prediction tools can shift power toward Big Tech and away from people.
In short: A new book by philosopher Carissa Véliz argues that generative AI is mainly a prediction tool, and that this helps Big Tech gain power while adding social risks.
Carissa Véliz’s book Prophecy: Prediction, Power, and the Fight for the Future, from Ancient Oracles to AI was published on April 21, 2026. In it, she says many modern technologies that look like they deliver “knowledge” are really built to make predictions, and those predictions can be used to control choices.
Véliz says this includes generative AI, the kind of AI that writes text or makes images. She points to how these systems often work by guessing what comes next, like predicting the next word in a sentence (similar to a phone keyboard suggesting your next word, just much bigger).
She argues predictions are not neutral. When predictions are used to judge people, like who gets a job, a loan, or extra police attention, they can lock in unfairness and make it hard to challenge decisions. She also warns of “self-fulfilling prophecies,” where people act as if a prediction is inevitable, and that behavior makes it come true.
Véliz has been promoting these ideas in recent appearances, including a TED2026 talk called “Beware the Power of Prediction” on April 14. She has also discussed prediction markets, privacy, and the limits of AI forecasting on the Big Technology Podcast.
Many AI tools are being sold as helpful assistants, but Véliz’s point is that a tool that mainly predicts can still be wrong while shaping real life outcomes. For regular people, the practical question is who benefits when predictions guide decisions, and whether there is a clear way to see, question, and appeal those decisions.
Source: NYTimes