The Cost of Progress: When AI Becomes a Weapon

by
Victoria d’Este

Printed: October 24, 2025 at 9:29 am Up to date: October 24, 2025 at 9:29 am

by Ana

Edited and fact-checked:
October 24, 2025 at 9:29 am

To enhance your local-language expertise, generally we make use of an auto-translation plugin. Please be aware auto-translation will not be correct, so learn unique article for exact data.

AI has change into part of our day by day lives. It helps us get work executed, sort out countless duties — and even suppose for us generally. However like two sides of the identical coin, it brings each advantages and dangers. That very same pleasant chat assistant can simply as simply be was a hacker’s software. Progress at all times comes with a value.

On this article, we check out how AI is reshaping cybercrime. Right this moment, assaults could be launched sooner and extra effectively than ever earlier than. Specialists from DeNet — leaders in decentralized safe storage — share their perspective on staying one step forward on this altering panorama.

Cybercrime on the Rise

Cybercrime has at all times been a problem, however lately it’s hitting more durable than ever — and the unfold of applied sciences like AI is one cause this pattern is accelerating. International losses from cybercrime are anticipated to succeed in $10.8 trillion by 2026 — up from $3 trillion in 2015, in response to Cybersecurity Enterprise. Each minute, criminals trigger large monetary injury, and the pattern exhibits it’s not going to decelerate.

“Latest assaults present that everybody is in danger: people, corporations, and governments. Ignoring your digital safety makes the implications inevitable — like skipping brushing your enamel after which being stunned by cavities,” mentioned Den Shelestov, co-founder at DeNet.

As cyber threats develop, many corporations are falling behind. A Darktrace survey discovered that 78% of safety leaders see AI-driven threats as an actual problem, and lots of don’t have the suitable abilities, data, or workers to reply successfully.

Hacking, Now on Steroids

Cybercrime is nearly as previous as computer systems themselves. Networks appeared, exploits adopted — and each new expertise is weaponized in file time. That’s the merciless regulation of progress: it creates alternatives for heroes and villains alike.

Hackers comply with a easy playbook: choose a goal, collect data, discover a approach in, transfer by way of the system, and seize information or demand a ransom. AI can help at each stage.

It used to take days, deep technical abilities, and a complete workforce to drag off a full assault. Now AI finds the holes, writes the exploits, and even pinpoints one of the best stress factors for ransom calls for, so even a child with a few scripts can assemble working malware.

However wait… aren’t LLMs like ChatGPT designed to dam malicious exercise?

In idea — sure. In observe — not likely. Hackers get round safeguards with immediate injections or jailbreaks, and a few even construct unrestricted LLMs. These black‑market fashions can generate malware, phishing campaigns, deepfakes, and extra.

How Hackers Are Weaponizing Know-how

Social engineering has at all times been a weak spot — and AI has now supercharged it. In keeping with SentinelOne, phishing assaults have jumped 1,265%. It’s hardly shocking, on condition that AI can craft a whole bunch of extremely personalised, convincing messages. You would possibly suppose a colleague shared a hyperlink to a undertaking replace — however one click on, and malware silently infects your system.

Deepfakes are one other rising risk. Advances in artificial audio and video have reached the purpose the place detecting fakes is extraordinarily troublesome. Attackers steal identities — high managers, executives, and shoppers — to trick staff into handing over cash, delicate information, or system entry. In 2024, Arup suffered large losses for precisely this cause: an worker was deceived throughout a faux Zoom name and ended up transferring $25 million to the attackers.

The subsequent frontier is agentic AI — autonomous methods that resolve and act with minimal oversight. In August 2025 Anthropic confirmed that their Claude mannequin was utilized in a fast, multi‑goal information‑extortion marketing campaign hitting no less than 17 organizations. Claude didn’t simply comply with directions — it ran the assault finish‑to‑finish: reconnaissance, exploitation, and bespoke ransom calculations — adapting at each step.

Assaults are sooner, smarter, and fewer forgiving. How will we cease one thing that may study and improvise on the fly?

What’s subsequent? 

The foundations of the sport have modified. Human groups alone can’t sustain anymore. That’s why corporations are preventing hearth with hearth, defensive AI is turning into the brand new frontline. Methods powered by machine studying can detect anomalies, reply to threats in actual time, and analyze patterns far past human capability. Nonetheless, even one of the best methods can’t cease somebody from clicking the incorrect hyperlink. Coaching stays important — don’t belief each e-mail, each name.

However there may be one other huge downside. Individuals and firms are unintentionally feeding the cybercrime machine. Large quantities of delicate information — emails, job titles and paperwork — are saved throughout organizations and third-party providers. A single leak could be mixed with different breaches and public data to assemble full profiles, making id theft, account takeovers, and extremely focused assaults far simpler.

“Knowledge leaks occur far too typically — it’s one thing we’ve been involved about for a very long time. Firms accumulate large quantities of delicate consumer information — which naturally makes them a golden goal for attackers. And there’s at all times the chance of a leak. We belief third events to retailer our information, however they’ll’t at all times assure its security,” Den tells.

That concern led to DeNet, a decentralized storage protocol designed to place management again in customers’ palms. Knowledge is encrypted on the consumer aspect, cut up into fragments, and distributed throughout a community of impartial nodes. Every fragment is saved on a number of units, eliminating single factors of failure. Solely the consumer’s personal key can reconstruct the complete information — shifting accountability and energy from companies to people.

Altering the perspective in direction of information storage is one option to struggle again towards cyber attackers — nevertheless it’s solely a part of the puzzle. Know-how retains evolving, new assault strategies consistently seem, and the battle between cybercrime and cybersecurity is endless. The one common rule is that this: your information has worth, and in the beginning, you might be liable for it. Shield it properly, and you’ll keep in management.

About The Writer

Victoria is a author on quite a lot of expertise subjects together with Web3.0, AI and cryptocurrencies. Her in depth expertise permits her to put in writing insightful articles for the broader viewers.

Extra articles

Victoria d’Este