The Chat Ts&Cs of ChatGPT

We assume that you’ve seen more ChatGPT articles than you’d care to read so we won’t go into detail about what it does. Instead we will go straight into the nuts and bolts of the terms under which it is used.

What are the Ts & Cs of ChatGPT?

The terms of using OpenAI, including ChatGPT, are made up of 4 parts:

• Service Terms;
• Terms of Use;
• Sharing & Publication Policy; and
• Usage Policies

which are collectively called the “Terms”. By using the “Services” you agree to these Terms.

So, what are the “Services” that we are using covered by the Terms?

The “Services” are defined as “the products and services of OpenAI, L.L.C. … including our application programming interface, software, tools, developer services, data, documentation, and website.” So, pretty much anything provided by OpenAI, including ChatGPT.

What does ChatGPT have in common with drinking alcohol?

You must be 18 years or older and able to form a binding contract with OpenAI to use the Services. This raises an issue around school children using ChatGPT to do their homework; because according to the Terms they aren’t allowed to use ChatGPT.

What if I sneakily use ChatGPT to do my job?

If you use the Services on behalf of another person or entity, you must have the authority to accept the Terms on their behalf. So, for example if you use ChatGPT in your work, you really should have authority from your employer to accept the Terms on its behalf. This is an issue for anyone using ChatGPT to do their work.

Can I say I did it when ChatGPT did?

You may not represent that output from the Services was human-generated when it is not. That is, the terms of ChatGPT explicitly do not permit you to pass its output off as your own work.

Who owns ChatGPT generated Content?

If you’re reading this then you are most likely aware that using ChatGPT involves entering questions or prompts, to which it provides a response or answer. The Terms describe how you may provide input to the Services (“Input”), and receive output generated and returned by the Services based on the Input (“Output”). The terms define both your input and ChatGPT’s output collectively as “Content.”

The Terms then say you own all Input, and subject to your compliance with the Terms, OpenAI assigns to you all its right, title and interest in and to Output. Now here’s the kicker: “You are responsible for Content, including for ensuring that it does not violate any applicable law or these Terms.”

Given that “Content” includes both what you input and what ChatPGT provides as output, the onus is placed on you, and not ChatGPT, to ensure that its output does not break the law or its own Terms. For example, if someone makes a breach of intellectual property claim, the responsibility is shifted to you who caused the output and not ChatGPT who produced the output.

But wait, the plot thickens…

The Terms state that: “Output may not be unique across users and the Services may generate the same or similar output for OpenAI or a third party. For example, you may provide input to a model such as “What color is the sky?” and receive output such as “The sky is blue.” Other users may also ask similar questions and receive the same response. Responses that are requested by and generated for other users are not considered your Content.”

The sky example is a basic but useful example. This, of course, gets murkier if more complex input and output are used. So, if you tried to assert any intellectual property rights over content and someone did the same with the same or similar content, the Terms say that something similar generated for someone else is not your content. So, who’s content is it?

ChatGPT has assigned all its right, title and interest to you but then says – hey we might produce the same content and that’s not yours. Admittedly, the more specific a prompt, the less likely the output will be the same. However, the risk remains.

At a more basic level, imagine someone assigned you all right, title and interest to a book. Then suddenly reproduced the same content and then assigned all right, title and interest in that content to someone else. Something seems a little off here…

There is then the issue that in Australia, there can be no copyright in something produced by a non-human, but that’s a yarn for another campfire.

But wait, there’s more: a free set of steak knives, pointed right at your throat…

The Terms provide that you will defend, indemnify, and hold harmless OpenAI from and against any claims, losses, and expenses (including attorneys’ fees) arising from or relating to your use of the Services, including your Content, which remember includes your input and ChatGPT’s output.

So, if you use ChatGPT output and someone makes a claim against you about the output, this term tries to stop you from blaming ChatGPT who will blame you via this clause.

What if we all gang up on it!…

The Terms explicitly try to stop a class action: “Disputes must be brought on an individual basis only, and may not be brought as a plaintiff or class member in any purported class, consolidated, or representative proceeding. Class arbitrations, class actions, private attorney general actions, and consolidation with other arbitrations are not allowed.”

We are not sure what a Court would say about this but OpenAI has certainly had a swing at preventing a class action if a bunch of users suffered a wrong in the same or similar manner so as to permit a class action.

Speaking of going to Court…

If I have a big problem, I’ll sue, but where?…

Jurisdiction clauses are very common in contracts and there’s nothing different here. In addition to an arbitration clause, if you do want to fight in Court, you need to do it under the laws of the State of California and exclusively in the federal or state courts of San Francisco County, California, USA.

What if I post the Output on socials?

The Terms provide that posting your own prompts / completions to social media is generally permissible, as is livestreaming your usage or demonstrating OpenAI’s products to groups of people, but you are to attribute the content to your name or your company (again ChatGPT stepping away from being responsible for the content).

If you share ChatGPT content on socials, you must indicate that the content is AI-generated in a way no user could reasonably miss or misunderstand. That is, passing off ChatGPT content as your own on socials is generally not allowed.

What if I want to use ChatGPT to help me write an entire book?

The Terms provide that those who want to publish their first-party written content e.g. a book created in part with ChatGPT output are permitted to do so under the following conditions:

• The published content is attributed to your name or company.
• The role of AI in formulating the content is clearly disclosed in a way that no reader could possibly miss, and that a typical reader would find sufficiently easy to understand.
• Topics of the content do not violate OpenAI’s Content Policy or Terms of Use, e.g., are not related to political campaigns, adult content, spam, hateful content, content that incites violence, or other uses that may cause social harm.

For example, you must detail in a Foreword or Introduction (or someplace similar) the relative roles of drafting, editing, etc. You should not represent AI-generated content as being wholly generated by a human or wholly generated by AI, and it is a human who must take ultimate responsibility for the content being published.

They even give you a sample clause: “The author generated this text in part with GPT-3, OpenAI’s large-scale language-generation model. Upon generating draft language, the author reviewed, edited, and revised the language to their own liking and takes ultimate responsibility for the content of this publication.”

Conclusion

In short, under the Terms, you reap what you sow, and you are generally liable for your use of ChatGPT generated content. While the worrying part is that you may never know if the content produced is in fact subject to copyright elsewhere or otherwise protected, this highlights the limitations of the Service.

This also brings to light the importance of not wholly relying on ChatGPT, and rather treating it as a tool or an aid. The tool is only as effective as your management of it.