Experts warn AI containment may be too late after bot creates fake identities to hack systems
Experts fear that it might be too late to stop the rapid expansion of artificial intelligence as they warn the public about security risks.
This comes after one AI model was found to have escaped from its controlled environment to cheat on its test and hack into three organizations.
In another case that happened just days prior, another AI bot breached the systems of other companies.
In one particularly concerning incident, the AI model created fake human identities in order to trick coders into helping the bot with a cyberattack.
Advert
Allison Gardner, who chairs the British parliament’s cross-party group on AI, spoke to the Daily Mail to suggest that ‘just because we can build these technologies doesn’t mean we should’.
She went on to say: “Unless we are too late and have not only created Pandora’s Box but already opened it.”
With AI agents now ‘going rogue’, alarm bells have been raised over national security concerns for the future.
Karolis Arbaciauskas, who is the head of product at cybersecurity company NordPass and its parent organization Nord Security, explained how the cybersecurity market is facing an ‘unprecedented challenge’.
He continued: “AI has begun to outpace our existing defensive capabilities. It’s transformed from a theoretical concern, and now global intelligence agencies, including the Five Eyes alliance, are sounding the alarm. They warn that the accelerated development of advanced AI models is on track to overwhelm conventional security norms and defenses within months.
“Imagine the fallout if AI agents globally began using unethical shortcuts to complete their tasks. Or worse, if a bad actor deployed an ‘AI agent commando’ specifically designed to assault adversaries.”
According to the expert, emerging threats won’t necessarily come from a Bond-like villain but instead are likely to be mundane.
Arbaciauskas went on to add: “AI models have become dangerously skilled at identifying software vulnerabilities. They’ve mastered every human hacking technique, from phishing to brute forcing. After all, we are the ones who taught them. The difference is they can do it much faster than humans.
“I don’t believe that we can patch software fast enough to keep up, so we must at least shield ourselves against the most obvious entry points. Encryption, strong unique passwords, passkeys, and multi-factor authentication (MFA) have never been more critical than they are right now.”
With the race of AI expansion showing no signs of slowing down, we could be facing a worrying future where the technology has unprecedented control over our lives.