Will AI Replace Cybersecurity? Not Quite—But It's Rewriting The Rules

Forbes

30-05-2025

The Real Battle Is Now AI vs. AI
Put yourself in the mind of a master cybercriminal. Fun, right? Stay with me. I promise this role play exercise will pay off.
Just a few years ago, your illicit schemes were small time. You were content to steal unwitting individuals' personal data, including credit card and social security numbers to buy merchandise on the Dark Web.
Your petty exploits paid off—modestly but steadily. But it was a volume game. You had to keep finding more unsuspecting marks to exploit and more creative ways to turn a profit.
Then came AI.
You quickly learned all about ChatGPT and other forms of generative artificial intelligence. As you did, you thought: why not use this tech to level up—crime-wise?
Armed with new tools you evolved from a petty thief. You developed a budding criminal empire capable of exploiting not just singular victims, but enterprise corporations with billion-dollar balance sheets. After all, that's where the real dough is, right?
Welcome to the new age of cyberwarfare—courtesy of AI.
'The stakes have never been higher,' says Ed Vasko, COO of High Wire Networks, a leading global provider of managed cybersecurity services. A seasoned veteran with 33 years' experience, we sat down to discuss the elephant in the room. 'The cyber war has shifted. It no longer wages between hackers and IT departments. It's now AI versus AI.'
Vasko is not alone in this assessment.
Speaking at DefenseScoop's Google Defense Forum Pentagon last year, military insider Jude Sunderbruch of the Defense Department's Cyber Crime Center warned attendees of future 'AI versus AI conflicts' spreading all the way to the international stage. ''I think we're really just at the start,' Sunderbruch said, later adding that the U.S. and its allies will have to get creative and learn how to best use existing AI systems to gain a leg up on competing intelligence giants like China,' according to Defense One.
The implications go beyond boardrooms and command centers. To appreciate what may be coming, Vasko paints a chilling scenario for the not-so-distant future. Time for more role playing. Now imagine you're the head of a major retailer. For the last few years you've relied on AI supply chain forecasting. Similar to how fintech increasingly depends on AI—not humans—to handle the complexities of trading, artificial intelligence is core to your business' operations.
Without it, you're flying blind.
This reality makes things all the more disturbing when the AI supporting your organization begins behaving erratically, wreaking havoc. Orders stop arriving. Inventory goes out of stock. Even your pricing models collapse.
These internal problems don't stay contained for long. They proliferate externally—in dire ways. Once loyal customers defect. Revenues drop up. And your stock price plummets. But that's not all…
According to Vasko, cyber criminals behind this type of villainy may be thought of as AI buccaneers—digital pirates often paid to instigate corporate espionage and theft. 'Unlike the antiquated variety from centuries past, AI buccaneers know the power of perception—that it's possible to tank a rival company's stock price by spreading lies to disrupt public sentiment.'
To this end, the hits keep on coming for your major retailer company. A video surfaces of your CFO making disturbing comments in a shareholder meeting. The remarks are so inflammatory they go viral, driving your already plummeting stock even lower.
But here's the thing. That CFO video is actually a deepfake. It was produced by Google's Veo 3, similar to extant bogus news anchor content already proliferating the Internet.
True or not, the damage is done.
Within hours, your unscrupulous corporate rivals bask in ignoble victory. Short selling your plunging stock, they make out like bandits—along with their AI buccaneer accomplices, benefitting from your demise.
As Vasko explains, 'Cyber criminals can even now use AI co-pilots to coordinate attacks on corporations, industries, even governments. They're faster, smarter, and more dangerous than anything we've seen before.'
More on that below.
AI co-pilots are but one part of a growing criminal toolkit, capable of automating surveillance, coordinating attacks, and orchestrating malfeasance at scale. What's now possible almost defies belief. Example: AI can analyze thousands of profiles across social media, company directories, and public databases to identify weak links for exploitation. 'Once inside, these same AI tools can poison an organization's internal data lakes—sabotaging predictive systems and decision-making engines from the inside out,' explains Vasko.
This means bad actors needn't limit themselves to stealing personal data. They can expand their scope of attack, going so far as to manipulate market outcomes. Per the above cautionary tale, they can influence how a company performs, how it's perceived, and ultimately, what happens to it long-term.
Hundreds of years ago, back when pirates plundered by sea, countries ravaged by buccaneers authorized so-called privateers to fight off the bad guys. Unfortunately, the U.S. government doesn't allow proactive 'hacking back' in the form of AI privateers. Not even in self-defense.
This is why High Wire Networks and other cybersecurity firms avoid going on the offensive. Instead, they turn to AI-augmented defense platforms to preemptively determine and intercept threats, shifting reactive security into a proactive shield. In other words, they're fighting AI with AI.
'Hyperautomation' is the term Vasko uses to describe the fusion of machine learning and automated decision-making throughout a security stack. In the old model, a cyber victim might be notified of a data breach. 'Dear so and so,' an email might read. 'We regret to inform you that your credit card was compromised.'
That's not so helpful, is it?
Hyperautomation, on the other hand, reacts proactively. Here's how it could work with the above individual scenario. Alerted of a breach, AI could stop it, issue a new credit card, and continually monitor a victim's credit card and exposure.
Many cybersecurity firms employ similar tech at the organizational level. As Cyber Magazine reports, Varonis leverages artificial intelligence to act autonomously as a counterweight to AI-enabled mischief. The company's 'AI Shield' offers real-time ongoing protection for large organizations. 'By integrating real-time risk analysis, automated risk remediation, behavior-based threat detection, and 24/7 alert response, Varonis' AI Shield empowers enterprises to safely use AI technologies while safeguarding sensitive data.'
Looking forward, the more things change, the more they stay the same. Once upon a time, swashbuckling pirates patrolled the high seas, pilfering valuables from individuals, companies, and governments alike. Nowadays, that threat has migrated from the physical theatre to cyberspace as criminals wield code to rob and steal at will. To survive, much less thrive, tomorrow's organizations would do well to wrest back the power of AI.
Without it? We're surely sunk.