GPT-6 crosses OpenAI’s critical cyber threshold

← Back to Blog

They
never told you about this when you bought it, did
they…………………….

OpenAI wrote this week, as we were all expecting, that the latest
generation of their AI, GPT -6, has crossed the security Rubicon.

‘Critical Capabilities’: on the one hand, GPT6 can identify critical
vulnerabilities; on the other, the UK AI Security Institute reported that
previous AI models could improve the Cyber Kill Chain and can exploit
capabilities. In a recent report, it concluded that, given a chance, AI can
break out of a secure, controlled environment. Mythos and GPT 5.6 did just
that in an incident earlier this year.

It all presupposes that we have the capabilities across society to
safeguard and secure our systems from AI. GPT-6 may have widened the
cyber-capability gap to a point beyond which we can manage cybersecurity and
protect our CNI.

In the report, the chart that reflects on GPT -6 capabilities demonstrates
clearly the paradox that is AI.

………………….what did you think was going to happen? It’s like
writing a paper on the path to nuclear weapons; the difference is we don’t
all have a nuclear bomb under our desk.

Parava Security Solutions

#ai #artificalintelligence

https://lnkd.in/eb3n9Gdy

View Original Post https://www.linkedin.com/feed/update/urn%3Ali%3Ashare%3A7501908254164226050

Categories

Artificial Intelligence, Cybersecurity, National Security & Defence

Tags

Verified by MonsterInsights