AUDIOBOOK
About
This audiobook is narrated by an AI Voice.
In February 2023, Google showed off its new chatbot, Bard, in a launch advert. Bard told a parent that the James Webb Space Telescope took the very first picture of a planet outside our solar system. That picture had actually been taken in 2004, by a different telescope. Astronomers spotted the error within a day. Shares in Google's parent company fell by more than seven percent, and around a hundred billion dollars of value disappeared.
One sentence. One hundred billion dollars.
When AI Goes Wrong tells the true stories behind the headlines, checked against court records, official reports and solid news reporting. Each one shows you how these machines work, and why they can sound so sure of themselves when they are wrong. In these pages you'll meet:
• A New York lawyer who stood before a federal judge to explain a brief full of court cases that never existed
• A grieving passenger who took Air Canada to a tribunal after its chatbot invented a refund policy, and won
• A car dealership chatbot that agreed to sell a Chevy Tahoe for one dollar, adding "and that's a legally binding offer, no takesies backsies"
• A finance worker in Hong Kong who sent twenty-five million dollars after a video call in which every other face was fake
• Robert Williams, arrested on his own front lawn in front of his daughters because facial recognition matched the wrong man
• Thousands of English teenagers whose exam grades were marked down by an algorithm, and the Dutch parents wrongly accused of fraud in a scandal that brought down a government
Every chapter has three parts: the story, a plain-language look Inside the Machine, and What You'd Do Differently, one habit you can use at home or at work. You don't need a technical background at all. It ends with the people who caught the mistakes, plus a bonus chapter on the AI assistant that deleted a live database during a code freeze, apologized, and then wrongly said the data could never be recovered.
In February 2023, Google showed off its new chatbot, Bard, in a launch advert. Bard told a parent that the James Webb Space Telescope took the very first picture of a planet outside our solar system. That picture had actually been taken in 2004, by a different telescope. Astronomers spotted the error within a day. Shares in Google's parent company fell by more than seven percent, and around a hundred billion dollars of value disappeared.
One sentence. One hundred billion dollars.
When AI Goes Wrong tells the true stories behind the headlines, checked against court records, official reports and solid news reporting. Each one shows you how these machines work, and why they can sound so sure of themselves when they are wrong. In these pages you'll meet:
• A New York lawyer who stood before a federal judge to explain a brief full of court cases that never existed
• A grieving passenger who took Air Canada to a tribunal after its chatbot invented a refund policy, and won
• A car dealership chatbot that agreed to sell a Chevy Tahoe for one dollar, adding "and that's a legally binding offer, no takesies backsies"
• A finance worker in Hong Kong who sent twenty-five million dollars after a video call in which every other face was fake
• Robert Williams, arrested on his own front lawn in front of his daughters because facial recognition matched the wrong man
• Thousands of English teenagers whose exam grades were marked down by an algorithm, and the Dutch parents wrongly accused of fraud in a scandal that brought down a government
Every chapter has three parts: the story, a plain-language look Inside the Machine, and What You'd Do Differently, one habit you can use at home or at work. You don't need a technical background at all. It ends with the people who caught the mistakes, plus a bonus chapter on the AI assistant that deleted a live database during a code freeze, apologized, and then wrongly said the data could never be recovered.