Timeline for Policy: Generative AI (e.g., ChatGPT) is banned
Current License: CC BY-SA 4.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 9, 2024 at 23:37 | comment | added | FNia | "decreasing the amount of bullshit that humans have to moderate": why not actually use AI to detect and flag the potential bullshit so humans can review them more quickly and easily? If the problem is moderating bad answers, then who cares how the good and verified answers are generated? | |
| Jan 22, 2024 at 14:06 | comment | added | Stack Exchange Broke The Law | @Summer-Sky I'm not sure if AI-generated answers thoroughly reviewed by humans are allowed. The rule was introduced with the intention of stopping spam that looks like good answers, not actual good answers. You might have to ask a moderator. However, reviewing the answer is as hard as writing one yourself! | |
| Jan 22, 2024 at 6:04 | comment | added | Summer-Sky | the simplicity of this chain of thought baffles me whenever I encounter this rule. It illuminates deeply down the simple ego of this "community". Assuming that a human will not verify the response created by AI makes the whole site pointless, anyone could have written "bullshit" answers before that. (and I refuse to believe that every answer is "verified" by mods, needs evidence or it didn't happen) .... Also funny how at the end SO came to offering AI themselves ... | |
| Apr 3, 2023 at 12:37 | comment | added | DubDub | Yes, I have a rule for my team that says update the documentation and they definitely always follow it. | |
| Mar 6, 2023 at 4:33 | comment | added | Dawood ibn Kareem | Yes, but a person who writes bad answers can improve as they learn more - and that's possibly more desirable than banning them outright. A person who copies ChatGPT answers is either going to continue, or stop. It's bad if they continue, so it's desirable to somehow make them stop. | |
| Jan 10, 2023 at 4:35 | comment | added | Anne Quinn | Let's compare to self-driving AI: If a human driver hits someone, you take that human off the road; if an AI driver hits someone, you have to take every car that uses that AI off the road, as they're all the same instance of the AI that hit someone. If one ChatGPT answer is bad enough to get it banned, then it should be banned across the board as it's the same entity. | |
| Dec 8, 2022 at 2:20 | comment | added | Stack Exchange Broke The Law | @KevinB I think anyone who attempts to copy-paste answers from a bot more than once should be suspended simply for wasting moderator time to the degree they do. Until they invent a bot which can create good answers. | |
| Dec 7, 2022 at 23:28 | comment | added | user400654 | They shouldn't, however they don't often immediately result in a 7 day suspension (and shouldn't) | |
| Dec 7, 2022 at 23:24 | comment | added | jscs | Hmm, maybe I made my point too obliquely, @KevinB. I agree 100% that the terrible answers from the bot should be deleted, and the users posting them should be sanctioned. But...the justification for this is not the bot itself: it's that the answers are terrible. And therefore it seems to me that the actual question is: why should equally terrible non-bot answers get a pass? | |
| Dec 7, 2022 at 17:13 | comment | added | user400654 | @jscs because the user who posted it posted 10 of them in an hour and they're all garbage. The amount of effort to clean up the mess is far higher than the effort it took to create it. More often than not, the user making the mess will be incorrectly rewarded for it by unsuspecting users thinking their answers are correct just because they were well "written". Ideally we'd solve this by throttling input in some way, however no such throttle exists currently. A temporary ban is a useful stopgap in the meantime. | |
| Dec 7, 2022 at 3:50 | comment | added | jscs | This is absolutely right: this machine generation should be disallowed not because of the means but because of the results. It also points towards a more fundamental and important question: why should a horrible misleading unhelpful answer be treated differently just because it was authored by a fleshy being instead of a silicon one? | |
| Dec 5, 2022 at 18:52 | history | answered | Stack Exchange Broke The Law | CC BY-SA 4.0 |