Don Schuerman, CTO and Head of Advertising, Pegasystems.
Each few years, an industry-shaking occasion jolts even essentially the most change‑averse organizations to search for from their street maps and ask an uncomfortable query: Are we nonetheless counting on yesterday’s options to struggle tomorrow’s battles?
The most recent warning shot comes from headlines round Mythos, a brand new AI mannequin from Anthropic reportedly so highly effective that its creators are slowing its rollout amid issues it may very well be weaponized by cybercriminals to probe, assault and infiltrate enterprise IT methods. However whether or not Mythos itself proves to be the tipping level nearly doesn’t matter. The sign is loud and clear: The stability of energy is shifting quickly towards attackers who can transfer at machine velocity.
For firms nonetheless counting on getting older legacy methods, I imagine that is extra than simply one other safety scare.
The brand new Y2K is already right here.
Many are already drawing comparisons to Y2K, and for good cause. Within the late Nineties, organizations scrambled to repair a seemingly small flaw of their code earlier than the calendar flipped to the yr 2000. The worry wasn’t theoretical. If methods failed, planes may very well be grounded, energy grids disrupted and monetary methods thrown into chaos. Corporations spent billions patching, testing and praying. We prevented catastrophe, however largely by duct‑taping the previous.
Quick ahead to immediately, and the parallels are placing. There are nonetheless tens of hundreds of organizations with active Lotus Notes installations within the wild. Main banks, governments and insurers proceed to depend on COBOL mainframes written a long time in the past. Then there’s Oracle Varieties, Home windows Servers, SAP EEC—the listing goes on.
Positive, these methods are steady, however stability shouldn’t be the identical factor as resilience. They have been by no means designed to function in a world the place attackers can use generative AI to investigate APIs, reverse‑engineer workflows or take a look at exploits hundreds of instances per minute.
That is what I’ve come to see because the uncomfortable reality: Legacy methods have grow to be the mushy underbelly of recent enterprises.
The instinctive response is acquainted. Patch the vulnerability, add one other safety layer and write compensating controls. In different phrases, plug the leaks with our fingers and hope the dam holds.
I do not imagine that strategy will work anymore. One‑off patching assumes a static risk mannequin. AI‑pushed attackers are something however static. They study, adapt and iterate quicker than any human safety staff can reply. Treating this like a conventional improve cycle is like bringing a squirt gun to a drone struggle.
What’s wanted now shouldn’t be incremental restore, however enterprise-wide reimagination. That doesn’t imply ripping every thing out in a single day. It means essentially rethinking how methods are designed, how work flows throughout the enterprise and the way people and machines collaborate. It means shifting the main target from “How can we shield this previous course of and squeeze a little bit extra life from it?” to “If we have been constructing this immediately, understanding what we all know now, what would it not appear like?”
Paradoxically, the identical AI revolution empowering attackers may also help flip protection into offense, if used the precise manner.
There’s a variety of buzz round “vibe coding,” the place generative AI spits out functions or scripts based mostly on a immediate. That’s thrilling, and in lots of instances genuinely helpful. It’s not stunning that the ability customers of AI coding are skilled engineers. Their expertise are nonetheless wanted to keep away from delicate errors, safety flaws or brittle logic that collapsed underneath actual‑world circumstances. With out the precise guardrails, total enterprise methods may be wiped out in seconds. When AI writes code quicker than groups can totally perceive or validate it, velocity turns into a legal responsibility.
However we are able to do higher than simply rewrite present methods quicker. AI may also help organizations ideate new working fashions, counsel higher workflows and redesign processes to harness the ability of brokers, all earlier than a single line of manufacturing code is written. AI can floor bottlenecks, establish redundant steps and counsel completely new methods of delivering outcomes that legacy architectures merely can’t help. It will possibly function not simply in code however in a visible language that enterprise and IT consultants can perceive.
In different phrases, we are able to use AI not simply to construct quicker, however to suppose higher.
Patching the previous gained’t shield the longer term.
That is the place legacy transformation can lastly break away from its repute as a sluggish multiyear slog. With the precise strategy, enterprises can extra readily discover modernization situations. They’ll strain‑take a look at concepts digitally earlier than committing tens of millions of {dollars}, they usually can modernize in a manner that reduces danger relatively than compounding it.
Sure, criminals are shifting quick with AI, however so can the remainder of us. Y2K taught us that ready till the final minute is dear and harmful. Patching the previous gained’t shield the longer term. The organizations that thrive within the age of AI would be the ones that cease asking find out how to protect legacy methods and begin asking find out how to outgrow them.
Forbes Technology Council is an invitation-only neighborhood for world-class CIOs, CTOs and expertise executives. Do I qualify?
:max_bytes(150000):strip_icc()/HDC-GettyImages-668641904-9179dc9fe60446d8b4d8a08fbffcf46d.jpg?w=600&resize=600,400&ssl=1 "What Happens When You Combine Turmeric and Black Pepper")
Recent Comments