Stack Overflow moderators go on strike after being instructed not to remove AI-generated content Options

[Christian J]

https://devclass.com/2023/06/05/stack-overf...erated-content/

A number of Stack Overflow moderators have declared a “general moderation strike” to protest about being instructed not to remove AI-generated content “outside of exceedingly narrow circumstances.”
...
“Moderators are no longer allowed to remove AI-generated answers on the basis of being AI-generated, outside of exceedingly narrow circumstances. This results in effectively permitting nearly all AI-generated answers to be freely posted, regardless of established community consensus on such content,” the striking mods said in an open letter.

...
The background is that StackOverflow appears to be struggling to make sense of the impact of AI tools like GitHub Copilot and ChatGPT on its question and answer site. Traffic has fallen. The company’s initial reaction was to ban use of ChatGPT to post Stack Overflow answers via a “temporary policy”, to the approval of the community
...
“Stack Overflow ran an analysis and the ChatGPT detection tools that moderators were previously using have an alarmingly high rate of false positives. Usage of these tools correlated to a dramatic upswing in suspensions of users ...

Maybe a sign of what's to come once the AI bots are let loose on the web in earnest?

[pandy]

Not only bots. Faked content created by people, both text and images, will increase, I think. I've seen amazing images made with these tools. I've tried ChatGPT, tried to make it create an image but I didn't succeed. Everything seemed to work and I was given an URL for the resulting image, but nothing was there. I don't think I did anything wrong, put it's possible.

I was really scared by something. You know the text editor I prefer and that it's programmable with its own scripting language? The program hasn't been updated for a very long time and the author doesn't seem to have any plans for it either, the reason is he has more profitable programs, I think. The userbase probably consists of hardcore oldtimers like me, don't think they make many new sales. The documentation is in help files and aren't available anywhere on the web AFAIK. Someone asked ChatGPT to write a script for it. It made some mistakes, but it knew the syntax and made a fairly good try. How on earth did it know how to do that? Have they really sucked in information to that level, help files for obscure and almost forgotten programs?

It seems AI can be very helpful though. I guess the problem will be to control it, to separate good use from harmful use. Some of my image programs use AI. But in one case I don't think it is AI, they just use the word for its buzz. It's, supposedly, used for noise reduction, but I don't see it's learning and the company has a huge database from testing cameras and lenses, so I think the data rather comes from there. It's amazing anyway. Noise is just sucked out without smudging the image (not much anyway), as is the case with traditional NR.

[Christian J]

Not only bots. Faked content create by people, both text and images, will increase, I think.

I was thinking of the improved ability to spam the web with more convincing content for less work. For example, you may push consensus in a certain direction in forum discussions, with the only give-away being that the accounts are new. Forum owners in turn can use an AI bot to make the forum look more active than it really is (and thus hopefully attract more real users). Marketing departments can make more real-looking user reviews for less money.

Real humans can certainly use it too, such as journalists or bloggers that don't have the energy to write new content every day. Expect to see even more absurd news articles than usual, when the journalists don't even have time to sanity-check their AI-generated texts.

The documentation is in help files and aren't available anywhere on the web AFAIK. Someone asked ChatGPT to write a script for it. It made some mistakes, but it knew the syntax and made a fairly good try. How on earth did it know how to do that? Have they really sucked in information to that level, help files for obscure and almost forgotten programs?

Sounds like it.

I guess the problem will be to control it, to separate good use from harmful use.

Rest assured that's NOT gonna happen.

Some of my image programs use AI. But in one case I don't think it is AI, they just use the word for its buzz. It's, supposedly, used for noise reduction, but I don't see it's learning and the company has a huge database from testing cameras and lenses, so I think the data rather comes from there.

A lot of the AI hype does sound like buzz. Speaking of photography and crappy AI, saw this link the other day: https://www.fotosidan.se/forum/threads/ai-f...11#post-2330242 (in Swedish).

[pandy]

Not only bots. Faked content create by people, both text and images, will increase, I think.

I was thinking of the improved ability to spam the web with more convincing content for less work. For example, you may push consensus in a certain direction in forum discussions, with the only give-away being that the accounts are new. Forum owners in turn can use an AI bot to make the forum look more active than it really is (and thus hopefully attract more real users). Marketing departments can make more real-looking user reviews for less money.

Real humans can certainly use it too, such as journalists or bloggers that don't have the energy to write new content every day. Expect to see even more absurd news articles than usual, when the journalists don't even have time to sanity-check their AI-generated texts.

They already do. And there a robots reading the news and acting as "hosts" between radio shows.

[Christian J]

They already do.

What I envision is something much worse than today, like a tsunami of fake posts drowning out everything else, everywhere. Then even if moderators manage to identify and ban the bots, other bot accounts will protest in defense of the banned ones. The few remaining real humans may also defend the banned bots, refusing to believe they are not humans too.

And there a robots reading the news and acting as "hosts" between radio shows.

Haven't heard about that, but do they just read the news (like that Max Headroom character), as opposed to creating the news themselves?

[pandy]

I don't know. I read an article about it, but I don't remember the details or where it was.

I read another article yesterday about a Norwegian scientist called Inga Strümke. She's a particle physicist and docent in AI and has just released a book called Maskiner som tänker (Machines that think). It sounded interesting and popular enough for me to read. But it doesn't seem to be available in Swedish yet (but has a Swedish title, so it should be on its way). Hopefully the library will have it as an e-book.

Interestingly she compares AI with the invention of the car - or the atom bomb.

[pandy]

The potential is probably bigger than we can imagine. You don't follow the war, but I try to. I'm fascinated by how big a role IT has, both IRL in form of drones and strategically online, mainly on Ukraine's part. It's a kind of war we've never seen before. Yet soldiers die in trenches and civilians are slaughtered the old fashioned way. It's like it plays out on two levels and that's kind of mind boggling. Almost like a computer game that has effects IRL.

[pandy]

What about this then? Meta has presented a new project they call Project 92 which they hope and expect will break Twitter. To do that they plan to hook up with celebs like Oprah Winfrey och - Dalai Lama. If they get Dalai Lama I give up on humanity. 😂

The omnipotens some people think they possess...