Our information environment is busier than ever, and it's never been harder to trust. Is that piece of health advice grounded in research? Is that statistic true or did an AI bot make it up? Is arguing with each other on social media a waste of time or digital democracy in action? AI is often blamed for presenting fiction as fact, with good reason: it is intrinsic to the design of large language models (LLMs) that they sometimes endorse falsehoods. But what if AI could help, not hinder, our search for facts? In this fascinating book,
Andreas Vlachos and Tom Stafford combine insights from computer science and psychology to offer some answers to this essential question. They explain the fundamentals of how LLMs work, how humans reason when faced with contradictory information, and how AI models can facilitate constructive dialog, helping us better engage with facts and with each other.
