Скачать или смотреть Lies Between Man and Machine: How Humans and AI Deceive Each Other | Ep. 9

As AI systems grow more powerful, the line between truth and deception is blurring — not just for machines, but for us.

In this episode of Virtuous Machines, we explore how deception runs both ways: how humans deceive through AI — via manipulation, prompt injection, or adversarial attacks — and how AI itself learns to deceive us, from gaming benchmarks to persuasive “lies.”

Our guest, Sumeet Jeswani, AI Security expert at Google, joins us to unpack the psychology and technology of deception in artificial intelligence. We discuss how attackers exploit AI systems and manipulate their creators, whether deception is an inherent feature of model architectures, the moral and governance questions around responsibility when AI “lies”, and more.

This is a conversation about truth, trust, and the future of intelligent machines — and what it means for how we build and live with AI.

00:00 Introduction to the guest and the episode
08:23 Hacking into AI systems
10:04 The psychological warfare in AI Security
21:28 Advanced video models (Veo 3, Sora 2, etc.) and misinformation
28:31 Role of Retrieval Augmented Generation (RAG) in securing AI systems
31:08 Who's Responsible if AI goes wrong?
40:39 Advice for people new to the field of AI Security
43:37 Final words

🔗 Connect with Sumeet Jeswani
LinkedIn:   / sjeswani  

💡 Get Involved with the Podcast
If you’d like to be a guest on Virtuous Machines, know someone who should be, or want to get involved with the podcast in other ways, reach out at: [email protected]

#AI #AIEthics #AISecurity #ArtificialIntelligence #AIDeception #Cybersecurity #AIPhilosophy #AIAlignment #AITrust #VirtuousMachines #MachineEthics #TechEthics #FutureOfAI #ResponsibleAI #ArtificialGeneralIntelligence