ChatGPT and similar programs - What is the official status?

[the-right-writer]

I'm unimpressed, and the program is not new. It is an old program with updates.

However, there are people who think it is going to revolutionize the way we work and live. It's not. On Upwork, it is going to muddle the field for clients because it is being used to hide the fact freelancers have no skills and are being deceptive. The result will be ChatGPT (I think this is the 3rd or 4th generation) generated freelancer proposals, profiles, and work. The product of AI is akin to poor writing from a non-native speaker of the language that must be edited or rewritten.

I want the moderators to clarify the use of the program through Upwork. It has been stated it must be disclosed. However, it has also been stated that the program pulls from other content, which puts it in the category of plagiarism. Plagiarism violates the Terms and is reason for permanent closure of an Upwork account. Also mentioned it was OK if you could prove it was original. At this point in time, there is no program to say, "Yes, this is original content." The current AI detectors are as bad as the writing.

1) Is it against the Terms to use ChatGPT? If not, under what terms can it be used?

2) There is no way to verify original work. Therefore, how can a freelancer "prove" it is original? There is no way currently. Since it cannot be proved, does that not make it against the Terms because of plagiarism?

3) Since the detector programs are wretched and unreliable, how can Upwork know it is plagiarism? No one else can tell, and it's not a secret.

4) Please do not give us another version of the previous statement. Although you are not saying, "If you use ChatGPT you will be banned per the Terms for plagiarism." by following through with previous statements, it certainly appears to be a violation until there is a detector program or plagiarism becomes acceptable.

Thank you.

[prestonhunter]

Discussions about ChatGPT are very new.

I have searched Upwork ToS.

https://www.upwork.com/legal

Upwork has no official status for ChatGPT.

But it is fine for you to advocate that Upwork adopt your suggestions for this.

For me, an important thing is tha clients in affected niches are wise and thoughtful about this topic, regardless of Upwork's official policy.

For example, if Upwork adopts a policy that states that Upwork freelancers can not use ChatGPT when submitting job proposals, that doesn't mean that clients won't receive proposals from freelancers who use ChatGPT to generate a job proposal. It would simply mean that when the client DOES receive such proposals, that the freelancer violated Upwork ToS.

I would never want any client to assume that all of the freelancers they encounter on Upwork are 100% honest, honorable and competent. Many are! But not all are. And Upwork has no way to guarantee that a client won't encounter a freelancer who isn't breaking the rules.

As always, it if is fine to talk about what changes Upwork SHOULD MAKE to improve.

And it is fine to talk about how to deal with Upwork the way IT IS NOW.

It's important to keep those things separate when we discuss the platform.

[the-right-writer]

I simply want a clarifying statement regarding the program's use.

[prestonhunter]

Would you be comfortable if Upwork simply had no policy, and each user was allowed to decide for themselves how they use it or not use it?

Upwork has no policy regarding the use of Jasper.

Upwork has no policy regarding the use of Bard AI.

Or ChatSonic.

Or LaMDA.

[the-right-writer]

Yes.

[the-right-writer]

I'm against plagiarism, but it exists now in different forms. It would be great if Upwork removed and punished all plagiarists, but that isn't going to happen. I think ChatGPT does use plagiarism, but so do many of the "cheat" programs.

[tlsanders]

Why do you say Upwork has no official position when Valeria clearly stated that use was prohibited unless disclosed? 

[prestonhunter]

re: "Why do you say Upwork has no official position when Valeria clearly stated that use was prohibited unless disclosed?"

I am not disputing Valeria's position of authority and her ability to speak on behalf of Upwork.

Perhaps it would be more accurate to say:

I regard Upwork has having no official policy about ChatGPT because it is an emerging topic and it is not referenced by Upwork ToS as recorded here: upwork.com/legal . But having said that, I do not believe that Upwork's in-practice policy is confined to what is found here: upwork.com/legal . I would say that Valerie has answered some questions in the Community Forum and that those statements can be understood broadly as "ad hoc policy" or "emerging in-practice policy."

[celgins]

I'm interested to hear any clarifications from Upwork Moderators, but I suspect it won't reveal much.

I suspect Upwork's reluctance to fully address the use of AI-generated content in an official capacity may be related to: 1) how to achieve it programmatically (using code/scripts); and 2) whether it's even feasible to try. Achieving 1 and 2 requires both human resources and system resources, and it all costs money.

The interesting thing is, those AI-powered tools that generate content will likely get better over time as developers continue to tweak their code. I think plagiarism will become more difficult to detect over the next 1-3 years, and we already see the unreliability of current AI-based detectors.

[prestonhunter]

Clark: I agree with you.

I would add:
I do not doubt that there are people who feel strongly that "using ChatGPT constitutes plagiarism."

Or, perhaps providing a more nuanced interpretation: "using ChatGPT constitutes plagiarism in many use casees."

But I don't think this is a universally held opinion.

I believe that if we asked many of the people using ChatGPT if what they are doing constitutes plagiarism, they would say no. And they wouldn't simply be lying in order to justify themselves. They would be telling the truth about how they really feel.

I really feel that this is an emerging topic, not "settled science" or "settled religion."

Serious people are discussing this question.

Perhaps a few years from now people will look back with amusement that we called using ChatGPT "plagiarism."

We may have another opinion about it entirely.

We may be better able to explain why using ChatGPT is NOT plagiarism, but is something DIFFERENT. Perhaps it will be something DIFFERENT than plagiarism, but still be something regarded as bad.

Similar to how libel and slander are both bad. But they're not the same thing.

Vulgarity and profanity are not the same thing. But they're both bad.

https://www.wired.com/story/chatgpt-college-university-plagiarism/

[celgins]

Yes sir... you provided a lot there to think about.

Perhaps a few years from now people will look back with amusement that we called using ChatGPT "plagiarism."

Exactly.

I thought about this the other day while reading through another post regarding ChatGPT and other text-generating AI-powered tools. As you stated above, what constitutes plagiarism with regards to these tools is an emerging topic, but I have heard some organizations actually encouraging the use of tools like ChatGPT and Jasper to increase productivity and reduce the amount of time it takes to crank out content (e.g., data, blog posts, articles, summaries, etc.)

Soon, folks may think you're nuts for not using ChatGPT and similar tools.

So, in the very near future, companies may begin asking candidates: "How versed are you with AI-tools like ChapGPT, Bard, and Jasper? We stay pretty busy, so we need you to keep on top of things, and it's much easier with one or more of these tools."

[25005175]

Right now, ChatGPT-3.5 is in something of a gold rush market, I think. The program is open to the public, officially, for R&D purposes. I predict that the program (and its competitors) will soon(ish) be pay-walled, once uptake has risen high enough that it is no longer a novelty and has become, for some industries, a necessity. The global free trial period has served its purpose for OpenAI's development team extremely well, but it is also must be prohibitively expensive to maintain such wanton free usage indefinitely.

[tlsanders]

This isn't the final version. It's data sources are very limited. The paid version will be an upgrade from this one.

[the-right-writer]

Seriously? We can't get a reply from the moderators?

[m_terrazas]

---

Jeanne H wrote:

Seriously? We can't get a reply from the moderators?

---

I have marked your post, to see if someone appears here.

[the-right-writer]

I sent them a nice private message to please address the questions. Maybe they will reply to you.

[usearchme]

They are only ever going to say don't use chat gpt obviously lol I saw it on another thread somewhere 

but as you point out above they will never know who is using it and who isn't as the detectors are not reliable, so there is little point adding something to the terms and conditions they cannot enforce

[tlsanders]

They are very, very busy making weird random determinations about when threads are no longer "productive."

[prestonhunter]

It's like issuing a statement about when a volcano will stop erupting.

This topic is in flux.

Job postings from clients wanting to hire freelancers who can help them implement ChatGPT are the fastest growing type of job on Upwork. ChatGPT is a already a significant source of revenue for Upwork, and has the potential to grow to be far more lucrative. But we have seen comments from some freelancers who want it banned completely from the site.

My recommendation to Upwork moderators would be to not comment on this subject at the present time.

[tlsanders]

I might agree with you if they hadn't already declared (falsely, as far as I can find) that using it without disclosure was a TOS violation.

[25005175]

They are calculating how many lightyears' long the pole must be before they touch it again. And looking for a way to hold it.

[the-right-writer]

Waiting for a response...

I did not begin this topic. I am asking for clarification of statements that have been made by an Upwork moderator.

[25005175]

Apparently missed this: February 10, Generative AI was added to the list of Freelance categories. New Generative AI Category Added 

[celgins]

I missed this too.

Since Upwork has added Generative AI as a category in the Marketplace and for Project Catalogs, it sounds like an admission of acceptance of all forms of generative AI--including material generated by chatbots like ChatGPT.

In other words, it appears that Upwork is leaving it up to clients to detect AI-generated content and determine what is real or fake.

Interesting.

[prestonhunter]

That sounds like a logical conclusion.

I don't think that one needs to look at acceptance of ChatGPT as an all-or-nothing proposition.

For example, a freelancer can work on Upwork even if they know that it allows ChatGPT, yet still choose to not use ChatGPT in their personal life. For example, an Upwork freelancer may choose to not use ChatGPT when they write a Christmas letter to their family and friends. An Upwork freelancer may choose to not use ChatGPT when they reply to a dinner invite on a dating website.

Upwork allows many things that I personally don't like. That is a strength.

[the-right-writer]

This does not explain the moderators post.

[kochubei_valeria]

Jeanne and others,

Thanks for your patience. Like these tools, our guidelines are evolving. At this time, Upwork does not expressly encourage or prohibit the use of AI; how you work and the tools you choose to use are for you and your clients to discuss.

Building a trusted relationship requires transparency and communication, which is why we recommend disclosing if you use generative AI as a tool in your work, including to help enhance productivity and project deliverables. Learn more