OpenAI releases new ChatGPT that's more ruthless and has fewer restrictions
Many are fearful that AI tools could 'break' the internet after a new release from OpenAI, as the latest ChatGPT model allows people to exploit tools that would previously be blocked by the company.
Hackers are constantly playing a cat and mouse game with cybersecurity experts, as while one exploit or vulnerability gets patched over and blocked, another one is bound to appear.
Security systems are more sophisticated right now than they've ever been, yet it's becoming increasingly difficult to protect against hacking attempts, especially now that bad actors have various AI tools at their disposal.
OpenAI CEO Sam Altman previously warned that there's likely to be a 'world-shaking' cyberattack as early as this year thanks to the availability of open source AI tools trained to exploit gaps within even the most sophisticated systems, and the company's latest tool could prove to be both a deterrent and the straw the breaks the camel's back.
Advert
As reported by the Independent, a new ChatGPT model has just launched specifically geared towards cybersecurity experts, as GPT-5.4-Cyber allows trusted organizations access to fewer guardrails and a more 'ruthless' approach.
The company has outlined that this model, similar to Anthropic's new Claude Mythos, is "purposely fine-tuned for additional cyber capabilities and with fewer capability restrictions," effectively giving the companies the same systems that hacker could be potentially using against them.
This, in theory, allows them to create the ideal barriers against the most sophisticated attacks that might come their way, and there's no better way of knowing how to defend than attacking yourself with the same tools as your enemy.
What might leave some fearful, however, is the chance – however small – that these tools fall into the wrong hands, as it's currently only accessible to those who pass through an internal vetting process.
OpenAI has expressed its desire to "make these tools as widely available as possible while preventing misuse," but it only takes one slip or leak for the floodgates to open — and Altman has previously expressed that its the responsibility of the world at large as opposed to individual companies to prevent the risk of any major cyberattacks.
That's far from the only major threat on the horizon though, as Google has also warned about the dangers that could arise from a 'quantum apocalypse', as quantum computing technology could shatter the security systems bound by encryption across the internet.
This, combined with AI, could completely eradicate every potential barricade in the modern digital era, so it's imperative that the proper safety precautions are in place and handled with caution before the danger actually appears.