discuss Who's Responsible When AI-Generated Code Goes Wrong?

[ActionAce]

I have used AI to generate a couple of codes especially when I'm lost and I can't seem to find a way out. I understand that it could be a double edge sword as AI generated code helps to boost productivity, it also raises questions about certainty. But when AI make mistakes, who do you think is to be blamed? Is it the developers, AI Creators or the AI itself?

 

[Zani]

Ai will keep having such issues because users' prompts differ. It is coded in a way that it will give different answers to different questions even when such questions are working towards a particular goal.

 

[Cpvr]

The ai system is to blame, however, it’s a user error at the same time. If you use ai for coding, you should always search regular programming help sites to help you troubleshoot the code.

 

[Oshawott]

In some countries, the laws states that the software owner is responsible. In some countries, those who developed the software is responsible.

As far as I am concerned, I think the software producer are responsible.

 

[Fidelia1]

The programmer is responsible if an AI generated code is wrong. The basic process of using codes to execute a program requires that codes should be debugged. It is the failure of the programmer to debug the AI generated code that causes the wrong code to be ran and he must take full responsibility.

 

[Shawn]

Probably not the AI or the maker of the AI.

This is because it's in their terms and agreements. If you don't read them, that's your fault, not theirs. They pretty transparent about AI not being 100% and that it can "hallucinate" from time to time.

Like others said, it's kind of your fault if you didn't check your work.

But using AI to code just adds extra work in that sense. You now have to check the work to make sure it will work. Before, you'd use AI to take away the step of coding but you still have to check your work now and that adds a step back to the process.

Logic tells me to just do it yourself in the first place.

 

[Cpvr]

In some countries, the laws states that the software owner is responsible. In some countries, those who developed the software is responsible.

As far as I am concerned, I think the software producer are responsible.

The prompts and input falls align with the user. It doesn’t fall back on the software producer if a user doesn’t check their own code. Stackoveflow and many coding forums exist for this very reason.

There are also many coding help sites around that will help you learn how to code, optimize the code and develop things without the use of AI.

 

[Yama]

It is the user. That's because if you check Gen AI tools, you can see a disclaimer "AI can make mistakes."

 

[lens]

We can't really blame one part at the expense of the other. I think it's a shared responsibility. It is the duty of the developer to review the code that is generated by AI. Also, the creator must ensure that the models that they rely on are transparent. Likewise, users must understand that generating a specific and clear prompt is important for getting better results.

 

[Zani]

It is the user. That's because if you check Gen AI tools, you can see a disclaimer "AI can make mistakes."

The type of prompts that users type in can go a long way in deciding a lot of things. So, I agree with you that the users should be blamed in such a situation.