Carissa Véliz, philosopher: ‘AI presents predictions as facts, and that has profound ethical implications’
Carissa Véliz, an Oxford University philosophy professor, argues in her new book *Prophecy* that AI-driven predictions are often treated as facts, shaping behavior and reinforcing power structures through probabilistic thinking. She traces the roots of prediction tools to colonialism and warns that such systems can manipulate expectations to alter reality, raising ethical concerns about their societal impact.
Carissa Véliz, a Spanish-Mexican philosopher and Oxford University professor, has released her new book *Prophecy*, examining how artificial intelligence and probabilistic predictions are reshaping power dynamics. In the work, she critiques the way AI presents predictions as facts, influencing human behavior and expectations in ways that can alter outcomes. Véliz highlights how this trend is tied to a broader culture of divination, where machine learning fosters a mindset that treats predictions as immutable truths. The philosopher argues predictions often function as disguised commands, shaping the world rather than merely describing it. She notes that probabilistic thinking, though mathematically recent, now underpins AI systems, influencing decisions from personal behavior to policy. Véliz warns that these predictions can create self-fulfilling prophecies, where expectations alter reality, particularly in human-centric contexts. Her analysis extends to historical roots, linking modern prediction tools to colonial-era practices like censuses and Francis Galton’s work on normal distributions. These methods, originally applied to gambling and astronomy, were later repurposed to impose normative standards on populations, labeling deviations as undesirable. Véliz emphasizes how AI amplifies these tendencies, embedding probabilistic frameworks into everyday decision-making. The book also explores why probability theory emerged late in history, suggesting cultural resistance to its compatibility with deterministic views of fate. She contrasts ancient Greek philosophy, which lacked probabilistic mathematics, with modern systems that rely on data-driven predictions to justify decisions. Véliz’s work urges a critical examination of how AI’s predictive capabilities are deployed, warning of ethical risks when predictions are treated as infallible directives rather than probabilistic estimates.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.