Artificial Intelligence didn’t just influence cybersecurity; it rewrote the rules overnight. What once required years of skill development can now be produced through a prompt in seconds. That shift created a new kind of skills crisis: not a shortage of people, but a surge of undertrained defenders who rely on AI outputs they don’t fully understand.
As someone who has lived both the offensive and defensive sides of cybersecurity, I see a widening gap forming. Attackers use AI to amplify existing knowledge, while too many new defenders are using AI as a substitute for building that knowledge. If we don’t fix that, no tool, no matter how advanced, will be enough to protect us.
AI Has Accelerated Cyber Offense Faster Than Cyber Defense Can Adapt
Attackers didn’t adopt AI to make life easier; they adopted it to scale attacks, improve precision, and reduce detection. Here’s how the threat has evolved:
Deepfake-Driven Phishing and Social Engineering
Higher-ed institutions are reporting an increase in AI-driven phishing campaigns featuring cloned voices, personalized lures, and synthetic identities designed to bypass traditional user awareness training.
Source: Ellucian
Polymorphic Malware and Code Evasion
Large Language Models can assist attackers by modifying malware signatures, rewriting payloads, and generating code variants built to avoid detection — a documented and growing offensive use case.
Source: arXiv
Attackers Are Getting Smarter, Not Lazier
The myth of “script-kiddies with AI” misses the real threat: capable adversaries who already know how to break systems, and now use AI to accelerate reconnaissance, pivot faster, and scale attacks.
Meanwhile, too many aspiring defenders begin by asking AI for answers instead of learning how cyber works. That’s a recipe for long-term failure.
The Emerging Gap: AI-First Defenders, Weak Fundamentals
We’re seeing a new trend: students and early-career professionals entering cybersecurity with an AI-first mindset. They learn how to prompt before they learn how to reason. That leads to dangerous patterns:
- Analysts who can run an AI-generated script, but can’t explain why it works
- Students who treat AI output as fact rather than a hypothesis to validate
- Defenders who freeze when a tool fails because they lack the core skill to pivot
AI didn’t reduce the need for fundamentals; it raised it. If defenders don’t understand core concepts, they can’t detect when AI is wrong, biased, or being manipulated by attackers.
AI should enhance capable defenders, not replace foundational knowledge they never built.
The Rise of “AI + Cybersecurity” Certificates: Helpful or Harmful?
The industry is now flooded with new training and certification programs branded as AI + Cybersecurity, promising fast-track mastery of AI-driven ethical hacking, prompt engineering, or automated defense strategies.
Here are some examples:
- AI for Cybersecurity Course (ISC²) — Foundational online course for cybersecurity professionals to understand AI threats and mitigation.
- Certificate in AI for Cybersecurity (Johns Hopkins University) — Self-paced certificate covering supervised/unsupervised learning, anomaly detection, biometrics, malware analysis, and deploying AI for defense.
- AI Fundamentals Certificate (ISACA) — Hybrid learning program covering core AI concepts, infrastructure, models and their security implications.
- CERT Artificial Intelligence for Cybersecurity Professional Certificate (CMU/SEI) — Focuses on the application, strengths and limitations of AI in cybersecurity.
These programs reflect the education industry’s pivot toward AI-powered threats and defenses. They are valuable, but they also carry a critical risk: if the emphasis is on tool usage and prompt mastery, without equal emphasis on core cyber skills, we are building a workforce vulnerable to failure when automation fails or attackers manipulate the machine.
Certifications must evolve to include, not replace, the basics.
AI in cyber education must be additive, not substitutive.
How Education Is Responding — and Where It Still Falls Short
Many schools and universities recognize the shift and are beginning to integrate AI into cyber programs.
Some positive moves include:
AI in University SOC Programs
Universities are using AI in campus Security Operations Centers to reduce alert fatigue, accelerate training of student analysts, and improve triage workflows.
Source: EdTech Magazine
AI-Integrated Cyber Curriculum
Some institutions now produce labs and coursework that teach AI-driven threat detection, response automation, and hands-on experimentation with AI tools.
Source: SCIRP
Teaching Safe AI Usage
Courses are increasingly including modules on the risks of entering sensitive data into public LLMs and proper handling of generative models in cyber contexts.
Source: ODU
Autonomous Pentesting Raises a Fundamental Question: What Skills Will Still Matter?
Platforms like Horizon3.ai and Pentera are reshaping how organizations test their security posture. Autonomous pentesting tools now can chain exploits, simulate attacker behavior, validate real risk, and deliver prioritized remediation guidance.
Example: Horizon3.ai
Which leads to an essential question:
If AI can execute full attack chains, do we still need human pentesters?
The answer is yes, but the role transforms.
Future defenders must:
- Validate and interpret automated findings
- Investigate beyond tool output
- Think creatively when automation misses the nuance
- Understand the why behind vulnerabilities and exploitability
The industry does not need a generation of tool operators.
It needs defenders who can outthink, interpret, and outperform the tools.
The Shift Cyber Education Must Make Now
If we want graduates who can defend in an AI-charged landscape, cyber education needs a reset.
Here’s what must change:
Fundamentals First — AI Second
Teach networking, OS internals, scripting, threat modelling, and adversarial thinking before introducing AI tools.
Teach AI Literacy, Not AI Dependency
Students must learn how AI works, where it fails, and how attackers exploit it.
Hands-On Adversarial Learning
Run scenarios in which AI gives wrong answers on purpose and train students to detect and respond.
Critical Thinking > Prompt Engineering
The future defender isn’t the one who writes the best prompt; it’s the one who can think when the AI is wrong.
Because if a student can’t defend without AI, they can’t defend with it either.
Final Thoughts: The Defender of Tomorrow
AI has changed cybersecurity; but it didn’t eliminate the need for human skill. It elevated the bar.
The defenders who thrive will not simply rely on AI to supply answers. They will be the ones who:
- Understand systems at their core
- Validate AI outputs
- Challenge the machine when needed
We stand at a crossroads.
We can build a generation that operates tools…
or we can build one that defends systems, with or without AI.
The future of cybersecurity depends on which path we choose.
About the Author
Len Noe is a Solutions Architect at BeyondTrust, Transhuman, Podcaster, International Cyber Security Speaker, Author, Technical Evangelist, and Biohacker with 11 implanted microchips. A former blackhat and ex-1% motorcycle club member with more than 30 years in technology, he has presented in over 70 countries and is featured in the documentary I Am Machine, which premiered at DEF CON 2025.