We're Becoming Dumb and Numb: Why the AI Hype Is Killing Cybersecurity (And Our Ability to Think)
0Random and Unscripted Episode: Black Hat 2025 AI Hype
"We're Becoming Dumb and Numb": Why Black Hat 2025's AI Hype Is Killing Cybersecurity (And Our Ability to Think) | Random and Unscripted with Sean and Marco
Sean and Marco expose Black Hat 2025's AI marketing problem on Random and Unscripted: every vendor claimed "agentic AI" while CISOs couldn't tell them apart. Using the Greek myth of Talos, they warn we're surrendering decision-making to imperfect systems, creating echo chambers that make us "dumb and numb" to reality.
Chapters:
00:00 Introduction and Icebreaker
00:15 Learning Language Quirks
01:10 Reflections on the Vegas Conference
02:30 Buzzwords and Marketing Challenges
04:47 The Role of AI in Cybersecurity
09:29 The Human Element in Decision Making
12:58 The Plateau of AI Innovation
22:44 Marketing's Responsibility and the Illusion of Perfection
25:18 Conclusion and Final Thoughts
Picture this: You walk into the world's largest cybersecurity conference, and every single vendor booth is screaming the same thing – "agentic AI." Different companies, different products, but somehow they all taste like the same marketing milkshake.
That's exactly what Sean Martin and Marco Ciappelli witnessed at Black Hat USA 2025, and their latest Random and Unscripted with Sean and Marco episode pulls no punches in exposing what's really happening behind the buzzwords.
"Marketing just took all the cool technology that each vendor had, put it in a blender and made a shake that just tastes the same," Marco reveals on Random and Unscripted with Sean and Marco, describing how the conference floor felt like one giant echo chamber where innovation got lost in translation.
But this isn't just another rant about marketing speak. The Random and Unscripted with Sean and Marco conversation takes a darker turn when Marco introduces the ancient Greek myth of Talos – a bronze giant powered by divine ichor who was tasked with autonomously defending Crete. Powerful, seemingly invincible, until one small vulnerability brought the entire system crashing down.
Sound familiar?
"Do you really wanna trust the weapon to just go and shoot everybody?" Marco asks, drawing parallels between ancient mythology and today's rush to hand over decision-making to AI systems we don't fully understand.
Sean, meanwhile, talked to frustrated CISOs throughout the event who shared a common complaint: "It's hard for them to sift through the noise and figure out what the heck is really going on." When every vendor claims their AI is autonomous and perfect, how do you choose? How do you even know what you're buying?
The real danger, they argue on Random and Unscripted with Sean and Marco, isn't just bad purchasing decisions. It's what happens when we stop thinking altogether.
"If we completely just use it for the easy button, we'll stop thinking and we won't use it as a tool to make our minds better," Sean warns. We risk settling for what he calls the "lowest common denominator" – a world where AI tells us what success looks like, and we never question whether we could do better.
Marco goes even further, describing a "circular knowledge learning" trap where "we use AI to create the knowledge, then to ask the question, then to give the answers." The result? "We're just becoming dumb and numb. More than dumb, but we become numb to everything else because we're just not thinking with our own head."
Their solution isn't to abandon AI – it's to get honest about what it can and can't do. "Stop looking for the easy button and stop selling the easy button," Marco urges vendors on Random and Unscripted with Sean and Marco. "Your product is probably as good as it is."
Sean adds: "Don't be afraid to share your blemishes, share your weaknesses. Share your gaps."
Because here's the thing CISOs know that vendors often forget: "CISOs are not stupid. They talk to each other. The truth will come out."
In an industry built on protecting against deception, maybe it's time to stop deceiving ourselves about what AI can actually deliver.