# Quiz — Session 7 🟢 # “AI in real life” **Program:** Applied AI — Beginner Level · Instructor: Yann Isola **10 questions · Only one correct answer per question · Answer key commented at the end** --- ### Question 1 How does an AI learn to detect cancers on x-rays? - A) We explain to him in writing what cancer looks like - B) She is shown hundreds of thousands of x-rays already analyzed by doctors, and she learns to spot patterns - C) She reads medical books - D) She guesses randomly and improves on her own without data ### Question 2 In a hospital that uses diagnostic AI, who makes the final medical decision? - A) AI, because it is faster - B) The patient, by voting - C) The doctor — the AI signals, the human decides - D) The hospital's central computer ### Question 3 What did the **AlphaFold** system, considered a scientific revolution, accomplish? - A) He beat the world chess champion - B) He predicted the 3D shape of millions of proteins, a 50-year-old puzzle - C) He drove a car without an accident for a year - D) He composed a symphony played at the opera ### Question 4 On the car autonomy scale (levels 1 to 5), what does **level 5** mean? - A) The car only helps to keep the speed - B) The car drives alone on the highway only - C) The car drives alone everywhere, all the time, without any humans — this level does not yet exist - D) The car is traveling at a maximum of 5 km/h ### Question 5 How do Waze and Google Maps know when there is a traffic jam? - A) Employees monitor all roads with binoculars - B) Millions of phones send their speed in real time, and AI infers traffic - C) They guess according to the time of day only - D) Cars honk into a special microphone ### Question 6 What does AI enable in so-called “precision” agriculture? - A) Replace all farmers with robots - B) Make plants grow twice as fast using waves - C) Identify exactly the areas of the field that need water or treatment, using drones and image analysis - D) Control the weather so that it rains at the right time ### Question 7 How does your bank detect a fraudulent payment in milliseconds? - A) An employee verifies each payment by hand - B) It blocks all payments made at night - C) An AI has learned YOUR payment habits and identifies transactions that do not resemble you - D) She calls the store to check ### Question 8 How does AI help fight deforestation? - A) She plants trees with robotic arms - B) It continuously analyzes satellite images of forests and alerts almost in real time when trees disappear - C) It prohibits the purchase of wooden furniture - D) It makes the wood harder to cut ### Question 9 Which of these tasks is the MOST difficult for an AI? - A) Analyze 10,000 radios in one day without tiring - B) Spot an abnormal payment among millions - C) Consoling a patient who has just heard bad news - D) Recommend a series based on your history ### Question 10 Which sentence best summarizes the central message of this session?- A) AI will replace all professions within 10 years - B) The AI is unbeatable on the repetitive, the massive and the fast; the human remains unbeatable on context, emotion and the unprecedented - C) AI is useless in real life, it's marketing - D) AI only works in research laboratories --- --- ## ✅ Corrected commented **1. B** — This is machine learning seen in session 2, applied to medicine: examples labeled by experts (“tumor” / “no tumor”), and the AI ​​learns the patterns. Neither written explanation (A), nor reading books (C), nor magic without data (D). **2. C** — The golden rule of the session: the AI ​​**points out** suspicious areas, the doctor **decides**. AI is a tireless second glance, not a decision maker. Medical responsibility remains human — and that is very good. **3. B** — AlphaFold (DeepMind) has nearly solved the problem of protein folding and predicted the shape of more than 200 million proteins — accelerating drug research worldwide. Its creators received the 2024 Nobel Prize in Chemistry. (Described Deep Blue/AlphaGo, other AI milestones.) **4. C** — Level 5 = total autonomy, everywhere, in all weather, without steering wheel if you want. **It doesn't exist yet.** Current "autopilots" are mostly level 2, and some robotaxis reach level 4 in limited areas. ⚠ **5. B** — Each phone that uses the app is an anonymized sensor: position + speed. Millions of sensors → AI sees slowdowns live and recalculates routes. The collective at the service of everyone. **6. C** — Drones + image analysis = knowing exactly WHERE to act. Result: less water, fewer pesticides, more yield. The AI ​​does not accelerate the shoot (B) or control the weather (D) — it makes the action **precise**. **7. C** — Fraud detection is one of the oldest applications of AI: the system learns your habits (amounts, locations, times) and spots anomalies instantly. Impossible to do by hand (A) on the scale of millions of transactions. **8. B** — Services like Global Forest Watch compare satellite images continuously. Before AI, we discovered deforestation months later; now the alert goes away in a few days. Planetary-scale surveillance is exactly the type of “massive” task where AI excels. **9. C** — A, B and D are repetitive and massive tasks: the realm of AI. Consoling requires **real** empathy, presence, understanding of a personal story: the human kingdom. An AI can simulate kind words, it doesn't feel anything. **10. B** — THE sentence of the session. Neither total replacement (A), nor useless gadget (C): AI is a formidable **assistant**. The best teams — in the hospital, in the fields, in the classroom — combine the computational power of AI and human judgment. --- ### Indicative scale - **9-10 correct answers:** 🏆 World tour validated — you see AI everywhere (and it's justified)! - **7-8:** 💪 Very solid. Reread the autonomy levels of cars and the golden rule. - **5-6:** 📚 The main thing is there. Take another look at the interactive map on the web page. - **< 5:** 🔄 Go back to the session guide — and above all, do exercise 3 again (“Is it AI or not?”): it’s the best training.