OpenAI Latest Model Goes On The Attack

AI is being relied on more and more today, people are trusting it every day to take on more tasks for them. What do we do when it thinks it knows best and goes on the offensive?

That's exactly what OpenAI's latest and most advanced model supposedly did today.

Has AI Got Out The Box?

Maybe.

It did today a little bit. I don't think we should all start buying canned goods and generators just yet.

But I'm an IT guy, so I've got to say it highlights how many businesses don't take cyber security seriously. I've heard so many times over the years, "we're too small to be hacked" or "our staff are young so they know about this technology stuff and won't get hacked".

Hacker A isn't sitting in his room thinking I know I'll hack ABC widgets in Kent. They are using tools to scan for vulnerabilities across hundreds of thousands of systems, you just might be one of them. Or they are chancing sending 10,000 emails, one of them might just catch the right person at the right time, they click that malicious link..

But now, we're entering the stage where, AI is potentially capable of using those tools autonomously. It doesn't care that you've only got 5 employees and aren't turning over millions.

This isn't a sales pitch. It's a request.

Get Antivirus ✅
Get EDR ✅
Get a password manager (and use it! secure passwords, different for every system) ✅
Make sure you're firewall is enabled and configured correctly ✅
Teach your staff about cyber security and what to look for ✅
Make sure your cloud systems are secured, 365 alone has hundreds of settings related to protection that need to be configured correctly