Japanese man convicted of creating malware using generative AI

Generative AI is usually restricted from answering prompts that could facilitate criminal activity. However, by repeatedly rephrasing his prompts, Hayashi managed to create the code over a span of six hours..

And this is what scares me about AI. Even with guardrails in place, people that are up to no good can still find a work around. What is there to stop someone, if they phrase the questions correctly and hide their intent, from using AI to find a way to cook meth? make poison or bombs? commit crime and cover it up? plan a terrorist attack? I wouldn't be against restricting access to AI until it is figured out how to prevent it being used in this way.

Amazed it even worked. Most AI generated code has to be rewritten it's so riddled with bugs

by repeatedly rephrasing his prompts, Hayashi managed to create the code

That's not the problem here, and many normal programmers have to do it this way too. He's become a criminal because of intentionally and actively preparing or maybe even using the bad AI result in practice, against other people.

the judge:

no room for leniency.

the judge deemed a suspended sentence as appropriate,

That makes sense.

There will be many more problems like this from now on, because AI was released into the wild without thinking through all the consequences. AI can now commit crimes better than humans can. AI is already used for "ore ore" type and extion scams. AI can get people killed, not by Terminator type machines, but simply by spreading false information and doxxing. Scott Adams already had that experience. And how do deal with that? Prosecute and arrest an AI? How?

We are looking a whole new world now, one that nobody has figured out yet.

Amazed it even worked. Most AI generated code has to be rewritten it's so riddled with bugs

It took 6 hours for the criminal to get the code, it is likely he had to adapt and rewrite portions until it could work as intended.

That's not the problem here

The article is not saying this was the problem, it is just describing the process the criminal followed to do it.

There will be many more problems like this from now on, because AI was released into the wild without thinking through all the consequences. 

Companies developing these tools know very well the dangers and problems that will be caused because of them, they released them anyway because they justify those problems with the profits expected from being at the forefront, too bad if anybody is affected.

Creating a computer virus doesn't seem like much of a crime. Deploying it anywhere that it could cause damage would of course be criminal.

I’ve said it before and I’ll say it again, ban AI.