ChatGPT fought the Law and the Law Won?

[jedishrfu]

In a recent article in Forbes, you could run into legal issues in using ChatGPT to further a business due to their indemnification clause:

https://www.forbes.com/sites/lancee...l-warns-ai-ethics-and-ai-law/?sh=1ff5df895719

 

[Borg]

So if GPT tells me to put rat poison in a cupcake recipe, I might be at fault?

 

[phyzguy]

So if GPT tells me to put rat poison in a cupcake recipe, I might be at fault?

Which is as it should be isn't it? If some link on the internet tells me to go shoot somebody and I do it, aren't I at fault?

 

[jedishrfu]

Yes and it can get worse than that!

I was thinking of all the writers, artists and others whose eork may have trained the model who lose their source of income due the GPT models and decide to sue. Case discovery may produce more litigants.

Also Forbes wrote another article on what you can and cant do with the GPTs.

https://www.forbes.com/sites/lancee...kly-prohibited-uses-per-ai-ethics-and-ai-law/

 

[256bits]

So if GPT tells me to put rat poison in a cupcake recipe, I might be at fault?

Just prove willing negligence and inferior product on their part, which will negate the indemnity clause.

 

[Vanadium 50]

In 1975
The Devil made me do it." (Geraldine Jones)
In 2023 - "ChatGPT made me do it"

 

[jedishrfu]

And ChatGPT will say to blame Geraldine Jones…

 

[Swamp Thing]

Also Forbes wrote another article on what you can and cant do with the GPTs.

https://www.forbes.com/sites/lancee...kly-prohibited-uses-per-ai-ethics-and-ai-law/

The article has a very long, rambling and repetitive preamble that almost turned me away. But finally he gets to a list of banned use cases. Many of these are obvious, but as the author points out, it would be possible to fall foul of them unintentionally because of the unpredictable nature of the technology.

 

[jedishrfu]

Yeah, he probably used ChatGPT to write the article saying I need 1000 words on what you are banned from doing.

 

[jedishrfu]

The Youtube lawyer Steve Lehto on ChatGPT legal exposure to users of ChatGPT:



This video seems to suggest that we shouldn't even identify posts as being generated by ChatGPT for fear of offending someone, and being held responsible for whoever sues ChatGPT.

 

[russ_watters]

This video seems to suggest that we shouldn't even identify posts as being generated by ChatGPT for fear of offending someone, and being held responsible for whoever sues ChatGPT.

I'm not hearing that. Do you have a specific timestamp/line? There's a lot of lawyer blablabla and a lot of letsthroweverycrapwiecanthinkofintotheEULAeventhougthecoutswon'tbuyit, but where the rubber meets the road, what matters is that a person who makes a post is responsible for the content of the post. I don't hear anything that says if we accuse a post of being chat-bot generated that the chat-bot owner or object of the post could sue us over it. It doesn't really even make sense to me.

 

[Tom.G]

I don't hear anything that says if we accuse a post of being chat-bot generated that the chat-bot owner or object of the post could sue us over it.

You are right, that concept is not in the video.

What IS in the video, and in the published Terms of Use which I read a week or so ago:
If you post something that ChatGPT generated and a third party sues the ChatGPT owner, then you, as a ChatGPT user/(the poster), must indemnify (pay back) ChatGPT for their costs of the law suit.

Here is part of Item 7 from https://openai.com/policies/terms-of-use

7. Indemnification; Disclaimer of Warranties; Limitations on Liability​

(a) Indemnity. You will defend, indemnify, and hold harmless us, our affiliates, and our personnel, from and against any claims, losses, and expenses (including attorneys’ fees) arising from or relating to your use of the Services, including your Content, products or services you develop or offer in connection with the Services, and your breach of these Terms or violation of applicable law.
...

Have a good day...
Tom

 

[jedishrfu]

The jist of the argument is:
- by using Chat you agree to the terms set forth in their contract
- you agree to indemnify Chat if your actions trigger a lawsuit
- some possible examples is if chat gave you as a response sime plagiarized text, posted it inline or used it in your own writing you could sued and so could Chat
- or if Chat said something untrue about a public person and you posted the untruth online you could be sued and so could Chat
- Chat would hire lawyers and you'd have to pay their bills and your own legal fees and whatever judgement was decided.

 

[russ_watters]

What IS in the video, and in the published Terms of Use which I read a week or so ago:
If you post something that ChatGPT generated and a third party sues the ChatGPT owner, then you, as a ChatGPT user/(the poster), must indemnify (pay back) ChatGPT for their costs of the law suit.

Right, that sounds perfectly logical/reasonable to me too. The person who makes the post is responsible for any harm it causes. Ideally the person harmed sues the person who made the post, but this covers the other direction; if they sue ChatGPT (because deeper pockets).

 

[russ_watters]

- you agree to indemnify Chat if your actions trigger a lawsuit
- some possible examples is if chat gave you as a response sime plagiarized text, posted it inline or used it in your own writing you could sued and so could Chat
- or if Chat said something untrue about a public person and you posted the untruth online you could be sued and so could Chat
- Chat would hire lawyers and you'd have to pay their bills and your own legal fees and whatever judgement was decided.

Does that seem harsh/unfair to you? It seems reasonable/fair to me.

Also, can you explain your prior comment about identifying chatGPT content? Who is "we" in the sentence: "we shouldn't even identify posts as being generated by ChatGPT"? (and why?)

60 Minutes had a surprisingly decent piece last night about AI stuff. In one part they were playing with Google Bard. I don't remember the specifics but they asked it for some information with references, and Bard gave nonexistent book/paper references, which they only found out later when they back-checked it. When they back-checked it. 60 Minutes was just playing, but they still recognized the need to back-check the output because they recognize that they are responsible for the content of their reporting.

 

[jedishrfu]

I was thinking that you post some self identified chatgpt generated content that some party objects to and sues you over it.

Because you say in the post that it came from chatgpt then the lawyers will think hmm this is a viable lawsuit with a good payout because we sue the poster and chatgpt for defamation.

You become the reason chatgpt got sued. By not mentioning where it came from then the lawyers will say the case isnt worth litigation because you alone have small pockets.

 

[Vanadium 50]

I'm just glad 60 Minutes has learned their lesson and is checking what they broadcast. Took them long enough!

 

[russ_watters]

I was thinking that you post some self identified chatgpt generated content that some party objects to and sues you over it.

Because you say in the post that it came from chatgpt then the lawyers will think hmm this is a viable lawsuit with a good payout because we sue the poster and chatgpt for defamation.

You become the reason chatgpt got sued. By not mentioning where it came from then the lawyers will say the case isnt worth litigation because you alone have small pockets.

Ok, I'll buy that at least in part: by identifying the source of the content you are identifying ChatGPT as a potential additional target. It's still a tough road with a catch-22 at the end. In addition to proving harm from the content and culpability by the software, they'd also have to successfully challenge the indemnity clause, otherwise the are left with the same problem - getting money from someone with shallow pockets.

 

[gmax137]

I'm just glad 60 Minutes has learned their lesson and is checking what they broadcast. Took them long enough!

Yeah, everytime they do a story that I have some knowledge of, I can see them being misleading. I'm not a regular viewer anymore.