Japan Today

Voices
in
Japan

quote of the day

I wanted to make money through ransomware. I thought I could do anything if I asked AI.

8 Comments

Ryuki Hayashi, an unemployed 25-year-old man from Kawasaki, who has been arrested for allegedly creating a computer virus by using interactive generative artificial intelligence available online.

© Yomiuri Shimbun

©2024 GPlusMedia Inc.

8 Comments
Login to comment

Another forewarning of what's likely to come.

5 ( +7 / -2 )

Ryuki Hayashi, an unemployed 25-year-old man from Kawasaki, who has been arrested for allegedly creating a computer virus by using interactive generative artificial intelligence available online.

he is thinking innovatively. There is a history of hackers being hired by government or industry to red team networks or software.

Given the narrowness of bureaucratic thinking I doubt that will be the case here.

2 ( +4 / -2 )

He's not real criminal, he just one unemployed person with no IT skill background was arrested because another newbie fraud case. No damage being done, there's no any victim. However the criminal who are really making virus still out there, not being arrested.

https://mainichi.jp/articles/20240528/k00/00m/040/101000c

-3 ( +2 / -5 )

I think this is going to be the next big ethical debate. How or should we monitor / restrict / report what people are asking AI to do? And who should do the monitoring / restricting / reporting? I don't have much experience with AI but I can see some nut case asking questions such at "How can I make an untraceable poison?" or "What is the easiest way to poison a barn full of rats?" (When they really mean office full of co-workers). "What materials are necessary to build a pipe bomb?" I know that info is already out there on dark web and such, but AI will make it even easier. Scary.

2 ( +2 / -0 )

Welcome to the future! One day, universities will start offering Bachelor of Science Major in Cyber Criminology, which could be accredited as a pre-Law course.

0 ( +0 / -0 )

Now give the AI the capacity to evolve its own ransomware for its own purposes. It would only need some kind of feedback on the effectiveness of each iteration. Suddenly we are talking a different thing.

1 ( +1 / -0 )

And who should do the monitoring / restricting / reporting?

Good question. BTW, those restrictions and prompting limitations are already massively applied, but can be and are obviously easily avoided by asking the same again and again until AI is somehow majority switched or triggered to answer although it doesn't answer those bad questions by original limited or restricted algorithm. The other possibility you have already named in your example, using formulations in the prompt that are less suspicious than explicitly or literally to ask for a bad question's answer from AI.

1 ( +1 / -0 )

Sven Asai

BTW, those restrictions and prompting limitations are already massively applied, but can be and are obviously easily avoided ... although it doesn't answer those bad questions by original limited or restricted algorithm.

Thank you for the information. I did not know this. I'm curious as to who comes up with the prompting limitations and restricted algorithms. I don't think there are government regulations. Are the companies basically self-policing?

0 ( +0 / -0 )

Login to leave a comment

Facebook users

Use your Facebook account to login or register with JapanToday. By doing so, you will also receive an email inviting you to receive our news alerts.

Facebook Connect

Login with your JapanToday account

User registration

Articles, Offers & Useful Resources

A mix of what's trending on our other sites