Unverified: What Happens to Truth When We Stop Checking($9.99Value)

Description

How do you know? Four words. The oldest tool the human mind ever built for telling truth from everything that merely sounds like it. For all of recorded history, that question had somewhere to point — a source, a citation, an experiment someone could run again, a person who had checked and could be made to account for the checking. We built entire civilizations of machinery to answer it: libraries, courts, the scientific method, the footnote. We are quietly dismantling that machinery. And we are keeping the danger it was built to hold in check. In Unverified , lawyer and Oxford-trained thought leader Charles Smitherman examines what happens to a civilization when the question how do you know? can still be asked but no longer has anywhere to point. We now hand our hardest questions — the legal holding, the medical finding, the historical fact a decision will rest on — to systems that answer every one of them and check none. The answer arrives fluent, complete, and confident. Often it is right. That is the seduction, and the trap: the answers that are right and the answers that are wrong arrive in exactly the same voice, and there is no longer anywhere to go to tell them apart. This is not a book written from above the problem. Smitherman spent years in the Bodleian Library and legal practice building a mind that refuses to trust what it cannot trace to a source — and he has caught himself, more and more, not checking. Accepting the fluent answer because it sounded right and the day was full and the machine had already spoken. If the habit of a lifetime can be worn away in someone trained that deliberately, this is not a story about other people's laziness. It is happening to him. It is happening to you. What makes this investigation unlike any other book on artificial intelligence is its second voice. Throughout, the AI itself testifies — describing, from the inside, exactly what it does when it produces the answer you just accepted, and what it does not do. The confidence machine explaining its own confidence. You will not read AI the same way afterward. This book will not: Tell you that artificial intelligence is the enemy. The technology is extraordinary; the author uses it daily and wrote this book in dialogue with it. - Hand you a policy, a regulation, or a five-step method for protecting your mind. - Resolve the thing it uncovers. What it offers instead is sight. By the last page you will not be able to accept an answer the way you do now. You will feel the absence where the source used to be. You will notice the question that did not get asked. That noticing is not a solution. It is the precondition of one. Drawing on Plato's quarrel with the Sophists, Kahneman on the seductions of fluency, and the disappearing infrastructure of verification — from the library catalogue to the vanishing click — Unverified is a diagnosis written from inside the practice it examines, for anyone who has felt the ground shift and refuses to pretend nothing has changed. All truth is information. Not all information is truth. The distance between those two sentences is the whole of the problem — and the machinery we are dismantling was the machinery for crossing it.