We make predictions all the time including about the weather, about traffic, about what someone is going to say next. It feels natural, even rational. But when algorithms start making predictions about us, whether we'll repay a loan, reoffend after prison, or respond to a medical treatment, something fundamental shifts. The forecast stops being a guess and starts becoming a verdict.
My guest today is Carissa Veliz, a philosopher and associate professor at the University of Oxford, where she also researches at the Oxford Internet Institute. Her work focuses on the ethics of technology, privacy, and artificial intelligence, and she advises companies and governments around the world on these issues. She's the author of the widely acclaimed book Privacy is Power, The Ethics of Privacy and Surveillance, and her new book, Prophecy: Prediction, Power, and the Fight for the Future, from Ancient Oracles to AI, is out now.
We talk about how the role of prophet has simply changed costumes throughout history from oracles and astrologers to economists and now tech executives and why that matters more than most people realize. Carissa explains how predictions about human beings are fundamentally different from predictions about the weather, why so many AI-driven forecasts are closer to commands than hypotheses, and what it actually looks like to take back your agency in a world increasingly shaped by algorithms.
Show Notes:
- [01:13] Carissa Veliz shares her background in philosophy, ethics, and advising companies and governments on technology and data.
- [02:35] She explains how prediction has existed throughout human history, from survival instincts to ancient prophecy.
- [03:49] The role of "prophets" evolves over time—from oracles and astrologers to economists, data scientists, and tech leaders.
- [07:05] Predictions about people differ from predictions about nature because they can influence outcomes and become self-fulfilling.
- [07:55] Many modern predictions, especially from tech leaders, function more like commands than neutral observations.
- [10:13] Carissa outlines key questions to ask when evaluating any prediction, including who benefits if it comes true.
- [10:13] She argues society has been overly naive about predictions, often mistaking power plays for objective knowledge.
- [14:18] AI systems are designed to please users, which can conflict with truth-seeking and scientific rigor.
- [14:54] Growing superstitions around AI include attributing agency, intention, or even spirituality to algorithms.
- [15:47] People begin trying to "please the algorithm," creating a modern version of superstition in digital systems.
- [19:55] The lack of regulation in AI places the burden of understanding risks entirely on individuals.
- [19:55] Carissa argues the real issue isn't just bias, but whether predictions about people should be used at all.
- [24:49] Insurance shifts from pooling risk across populations to targeting individuals, increasing inequality and personal burden.
- [27:02] Self-fulfilling prophecies in medicine and decision-making can hide their own failures by erasing alternative outcomes.
- [30:25] Predictive systems risk limiting human potential by filtering out those who don't fit expected patterns.
- [30:25] Society thrives when individuals can defy expectations, something prediction-heavy systems may suppress.
- [35:21] Algorithms reduce exposure to randomness, while real-world interactions create unexpected opportunities and insight.
- [36:11] Over-reliance on AI can replace human relationships and narrow life experiences.
- [36:11] Carissa reframes uncertainty as a positive force that enables freedom, choice, and democratic possibility.
- [36:11] She encourages treating predictions as possibilities to question—not instructions to follow.
Thanks for joining us on Easy Prey. Be sure to subscribe to our podcast on iTunes and leave a nice review.
Links and Resources:
