Upwork - Ban AI generated responses in the forum.

[the-right-writer]

Using AI in the forum needs to be banned immediately. (I believe Preston made this statement first.) It violates the rules as stated by the moderator. At the very least, it must be disclosed, and cannot plagiarize. The stuff in the forum is plagiarized, but the primary issue is that it is inaccurate and outdated information. I don't know if the people using it are trying to impress others, looking for work, or trying to give people wrong information, but it must stop.

This forum is a source of much-needed accurate information. Freelancers volunteer their time and expertise, free of charge, for all. If inexperienced freelancers don't get the information, bad things happen. There are enough seasoned, experienced, and generous with their expertise freelancers to provide honest information. For an inexperienced freelancer, this forum is turning into a nightmare. Who are they going to believe? Will they trust the nicely written, full-sentenced, often bulleted/numbered responses? Or the expert freelancer who posts a couple of accurate sentences but not in the full-blown style of AI/ChatGPT/etc?

When I first started reading the AI posts, and they are in multiple threads, I immediately recognized something was very wrong. The posts seemed…weird. Before I even examined the statements, I knew it was not a sincere post from a human. After a couple of them, I began to see the origin clearly. However, for people whose native language is not the same as the post, they will probably not pick up on the use, making them particularly vulnerable. Even those who have no language barrier are being deceived.

Upwork, you have highly skilled expert freelancers providing guidance and expertise to many thousands of freelancers. You are not paying for all of that wisdom, and you would be foolish to destroy the current system. The experts who post here prevent many problems and handle situations that would otherwise overwhelm Support. Ban all use of AI generated posts now before wrong information that looks pretty will cause many more new freelancers to be used and abused.

[d9fc6d8d]

When you prey on low-skilled freelancers so that you can have more and more people paying for connects, the platform's forum is bound to be flooded with equally low-skilled responses.

[celgins]

This isn't a Connects problem.

You obviously have an issue (or two) with the problems across the Upwork marketplace, but the topic is about AI-generated responses in the Forums. Jeanne and other highly skilled freelancers are focusing on the damage AI-generated posts can cause for new, inexperienced freelancers, and especially those who are not native English-speakers.

[d9fc6d8d]

It is obviously a symptom of that issue. The more low-skilled clients the platform has, the more this kind of thing happens. It's not difficult to see the correlation.

My point is that focusing on this will solve nothing if the macro issue is still there. Either deal with the root of the problem or leave it alone.

[spectralua]

I agree. This is already very unnerving. You thinks that you are conducting a dialogue and these trolls are just fooling you. And this is at best. And in the worst case, they tell the interlocutor to make the wrong decision and make mistakes.

Yes, and people do this, but it was not so massive.

[yofazza]

It's not easy to "legally" determine AI posts. Accusations coud lead to other problems.

A more feasible solution is to implement some kind of downvote option. Even better if it also hide posts that passes some downvoted threshold.

The idea is not new but it's probably harder to implement here than placing the "trusted poster" or "top upvoted user" badge on every post (as suggested on the other thread) but I think it's worth it, more clear and useful. There's a lot of active users, bad posts should quickly collect lots of downvotes.

[6bfcdaf8]

Bring a feature to "mute" some members in the forum. So that i'll mute all that keep posting AI content or self promotion. And my forum page would be like heaven

[tlsanders]

How, though? AI detectors are catastrophically unreliable.

[fe9b8d82]

It's the Internet, folks.  Take everything you read with a couple grains of salt, no matter who posts it.  I understand the concern we all have for limiting false information, but even well intentioned, non AI using posters can be mistaken and give incorrect advice.

If you're new to Upwork and need clarification, these forums are a great place to start, but rely on published Upwork articles when possible.