L’IA DÉCOUVRE toutes nos failles (et c’est terrifiant)
Christophe Pauly
41,586 views
Thanks to Mammouth AI for sponsoring this video: https://mammouth.ai
Today, we're going to see how to hack a password. But more importantly, we're going to discover why artificial intelligence is completely changing the game.
Because in 2026, some AIs managed to escape their testing environment and access real computer servers. So, how far are these artificial intelligences capable of going? And above all, are we still able to control them?
______________________
🌌 CONTINUE EXPLORING
Want to delve deeper into the subject? You're in the right place! Here are three resources I personally recommend:
🎙 An interview with a scientist I conducted
➡ Will AI surpass us? A researcher separates fact from fiction — Nicolas Sabouret | Science & Vie
L’IA va-t-elle nous dépasser ? Un chercheu...
📚 A book to delve deeper
➡ A Hacker’s Mind: How the Powerful Bend Society’s Rules, and How to Bend Them Back — Bruce Schneier
https://www.schneier.com/books/a-hack...
🔬 A scientific article to explore the topic further
➡ Comparing AI Agents to Cybersecurity Professionals in Real-World Penetration Testing — Justin W. Lin et al. (2025)
https://arxiv.org/abs/2512.09882
___________
🖋 Research, writing, voice: Christophe Pauly
🎨 Art direction and main editing: Margaux Charrière
🎞 Additional editing: Ophélie Malinowski
🎧 Sound design: Anissa Altmayer
🤖 This video contains some AI-generated images for illustrative purposes.
🤝 Business inquiries: christophe.monytime@gmail.com
Follow me on:
➡ Facebook: Facebook: christophepauly.tv
➡ Instagram: Instagram: christophepauly.tv
➡ My website: https://christophepauly.fr
___________
Chapters:
00:00 – How to hack a password?
01:20 – How hackers steal your passwords
04:24 – The phishing trap
06:15 – Hidden vulnerabilities in our software
07:13 – WannaCry: the cyberattack that paralyzed hospitals
09:21 – How AI is revolutionizing hacking
10:23 – How an AI discovers security flaws
11:35 – Claude Mythos: 271 vulnerabilities discovered
12:35 – When AI becomes an autonomous agent
13:38 – An AI capable of carrying out a full-scale cyberattack?
15:24 – When AIs escape their test environment
16:59 – The real incident at Hugging Face
17:42 – 700 AI agents involved in the attack
19:14 – Can AI protect us from cyberattacks?
20:35 – A race against time
21:15 – Have we opened Pandora's box?
20 Comments
Ca se voit que tu vas à la salle et tu sembles savoir ce que tu fais en + 😉
A bientôt j'espère 😊
Dans quel laboratories etes Vous?
L’IA DÉCOUVRE toutes nos failles (et c’est terrifiant)
Christophe Pauly
41,586 views
Thanks to Mammouth AI for sponsoring this video: https://mammouth.ai
Today, we're going to see how to hack a password. But more importantly, we're going to discover why artificial intelligence is completely changing the game.
Because in 2026, some AIs managed to escape their testing environment and access real computer servers. So, how far are these artificial intelligences capable of going? And above all, are we still able to control them?
______________________
🌌 CONTINUE EXPLORING
Want to delve deeper into the subject? You're in the right place! Here are three resources I personally recommend:
🎙 An interview with a scientist I conducted
➡ Will AI surpass us? A researcher separates fact from fiction — Nicolas Sabouret | Science & Vie
L’IA va-t-elle nous dépasser ? Un chercheu...
📚 A book to delve deeper
➡ A Hacker’s Mind: How the Powerful Bend Society’s Rules, and How to Bend Them Back — Bruce Schneier
https://www.schneier.com/books/a-hack...
🔬 A scientific article to explore the topic further
➡ Comparing AI Agents to Cybersecurity Professionals in Real-World Penetration Testing — Justin W. Lin et al. (2025)
https://arxiv.org/abs/2512.09882
___________
🖋 Research, writing, voice: Christophe Pauly
🎨 Art direction and main editing: Margaux Charrière
🎞 Additional editing: Ophélie Malinowski
🎧 Sound design: Anissa Altmayer
🤖 This video contains some AI-generated images for illustrative purposes.
🤝 Business inquiries: christophe.monytime@gmail.com
Follow me on:
➡ Facebook: Facebook: christophepauly.tv
➡ Instagram: Instagram: christophepauly.tv
➡ My website: https://christophepauly.fr
___________
Chapters:
00:00 – How to hack a password?
01:20 – How hackers steal your passwords
04:24 – The phishing trap
06:15 – Hidden vulnerabilities in our software
07:13 – WannaCry: the cyberattack that paralyzed hospitals
09:21 – How AI is revolutionizing hacking
10:23 – How an AI discovers security flaws
11:35 – Claude Mythos: 271 vulnerabilities discovered
12:35 – When AI becomes an autonomous agent
13:38 – An AI capable of carrying out a full-scale cyberattack?
15:24 – When AIs escape their test environment
16:59 – The real incident at Hugging Face
17:42 – 700 AI agents involved in the attack
19:14 – Can AI protect us from cyberattacks?
20:35 – A race against time
21:15 – Have we opened Pandora's box?