Nemstar Insights

Vibe Hacking, Claude, and the AI Shift: Hackers Are Using AI Now. Are You Ready?

By Sean Hanna, Founder & Director of Nemstar

February 16, 2025

Hackers are using AI. Right now. Not next year. Not five years down the line. Today.

Late yesterday, we got confirmation of what many of us in cyber suspected: Chinese state-backed attackers used Anthropic’s AI model Claude to automate the lion’s share of a recent cyberattack campaign. According to The Verge, up to 90% of the attack chain was driven by AI. Reconnaissance. Credential harvesting. Intrusion. All done with little human input.

This isn’t speculative. It’s not theoretical. This is what cyberattacks look like now.

So, let me ask you plainly:

If attackers are using AI to scale and automate their efforts, and you’re still defending with spreadsheets and outdated training modules… how do you expect to keep up?

Vibe Hacking: The Human Exploit

You’ve probably heard the term “vibe hacking” floating around. Sounds fluffy, doesn’t it? But don’t let the name fool you. This is the evolution of social engineering, powered by agentic AI. We’re talking about dynamic manipulation based on real-time emotional cues, context shifting, and decision trees that change based on you, your actions, your tone, your response.

It’s like phishing, but with a psychology degree and a limitless attention span.

We used to train people to spot dodgy emails. Now we need to prepare them for persuasive AI chatbots trained on human dialogue, capable of mimicking empathy and building rapport, all with the intent to breach.

The Playing Field Has Tilted

I’ve been saying it for years: “The difference between knowledge and understanding is action.”

Knowing AI is out there is one thing. But if you’re not acting, upskilling, adapting, building resilience, then you’re not ready. Full stop.

AI in the hands of attackers means:

 

  • Attacks are faster, cheaper, and harder to trace.
  • The skill barrier has dropped. Anyone can run advanced campaigns with a prompt and a stolen model.
  • Attack volume is going up. Your perimeter is going to get tested, daily.

 

It’s not a question of if. It’s when.

So What Do We Do About It?

At Nemstar, we’ve launched the Nemstar Certified AI Practitioner (NCAP) course precisely for this reason. Because we don’t just need to defend against AI, we need to leverage it.

This isn’t a buzzword course. It’s hands-on, structured around four key pillars:

 

  • AI Unboxed: Understand how LLMs and agentic AI work (and fail).
  • AI Governed: Build an AI risk register and map it to real-world frameworks.
  • Evil AI: Explore how attackers weaponise AI, prompt injections, jailbreaks, ShellGPT, WormGPT.
  • Good AI: Learn to apply AI in defence—incident response, behavioural monitoring, control tuning.

 

Because if they’ve got Claude, we need to have a counter.

Final Word: Don’t Wait to Be the Headline

There’s no excuse anymore. The attackers are using AI, right now. If you’re not preparing your team, your business, or yourself, you’re just waiting to be breached.

We’ve spent years training defenders. We know what works. And right now, the smartest thing you can do is invest in understanding AI as a risk and as a tool.

The NCAP course launches this January.

Are you going to be ready?

Coming soon


About the Author:

Sean Hanna

Founder & Director

Sean Hanna founded Nemstar in 2009, leveraging 20+ years of cybersecurity expertise to deliver business-focused technical training. As EC Council’s Global Security Trainer of the Year, he has trained thousands globally, including at Microsoft, GCHQ, and the U.S. Navy. Sean led Microsoft Exchange 2000’s technical training launch and helps organizations strengthen their cybersecurity defenses.

Explore Other Courses

Want to find out more about CRISC?