Everyday AI: what you see on screen, and what gets sent

You ask your phone for directions. An answer in a second. But in that second, what did you give away? In 1950, Alan Turing published a paper asking whether a machine can think, and proposed an “imitation game” to judge it. In 1997, IBM's Deep Blue computer beat world chess champion Garry Kasparov, in a match of six games. In 2012, a neural network trained on graphics cards won the ImageNet image contest, clearly ahead of its competitors. Three steps, 150 seconds. Step one: one gesture, two sides. On screen, an instant answer. Behind the scenes, out go your location, the time, your history. The benefit is real: the route adapts to traffic. So is the cost: your journey becomes data. The real question: can you see, then delete, what was collected? Step two: why now? AI dates back to the 1950s. But in the 2010s, computing power and huge amounts of data finally came together. Deep neural networks took off. In France, AI use in firms with 10 to 249 staff nearly tripled: 6% in 2021, 17% in 2025. Step three: two challenges. First, bias: fed unbalanced data, a system can repeat unfair decisions. Then the black box: even its makers can't always explain an answer. Harmless for a photo filter, not for health or justice. The EU AI Act tackles part of this: since August 2026, when you talk to an AI, you must be told. In the KOAEE assistant, at the town hall, the clinic or the insurer, one line stays below the chat: Koaee is an AI and can make mistakes. And on a risky question, it doesn't play expert: it refers you to a professional. To understand AI in 150 seconds, subscribe to the channel. The AI that stays between you. Try it at koaee.ai.