| ▲ | arrsingh 9 hours ago |
| There should be a "flag as AI" link in addition to "flag" and then a setting for people to show flagged as AI. Once the flagged as AI reaches a certain threshold then it disappears unless you enable "Show AI". Maybe once enough posts have been flagged like that then that corpus could be used to train an AI to automatically detect content generated by AI. That would be cool. Maybe the HN site wouldn't add this feature but if someone wrote a client then maybe it could be added there. |
|
| ▲ | dang 9 hours ago | parent | next [-] |
| We're going to add that. I've resisted adding reasons-for-flagging for years, but even I can change my mind every decade or so. A nice side effect is that it will double as a confirmation step, solving the FFF (fat finger flagging) problem. |
| |
| ▲ | altairprime 22 minutes ago | parent | next [-] | | > it will double as a confirmation step, solving the FFF (fat finger flagging) problem Thank you!!! | |
| ▲ | ninjagoo 3 hours ago | parent | prev | next [-] | | Will there be a process or opportunity for mis-flagged comments' posters to prove their comment was human generated? Or will they have to simply eat the karma hit and move on? | | |
| ▲ | dang 2 hours ago | parent | next [-] | | Anyone can email hn@ycombinator.com and ask us to take a look either way. | |
| ▲ | oblio 2 hours ago | parent | prev [-] | | Annoyingly as downvoting is, it's limited to -4. |
| |
| ▲ | mikewarot 7 hours ago | parent | prev | next [-] | | My radical opinion is there shouldn't be 2 flags, there should be N flags, user defined, so that we can flag humor/satire/factuality/insight/political and a bunch of other things. I fully realize that's not going to fly any time soon. Adding AI in addition to the standard up/downvote and flag seems a reasonable thing. | | |
| ▲ | tptacek 6 hours ago | parent [-] | | Flags are a signal to the moderation system. What does it mean to "flag" something as "factuality" or "satire"? | | |
| |
| ▲ | DetroitThrow 6 hours ago | parent | prev [-] | | Flag as AI would be incredible and is probably unique to software-focused forums. Saves everyone who wants it a lot of time. Still allows cool content to reach the front page with some visibility or escape some moderation queue. Thanks for not standing still on this issue. The world is changing, fast, and glad HN responded quicker than some forums on a cogent stance. |
|
|
| ▲ | altairprime 9 hours ago | parent | prev | next [-] |
| ‘Flag’ is an algorithmic flag only, and there are no humans in the flag algorithm’s processing loop. They may monitor and react to the ‘queue’ of flagged articles, and they can do special mod things with flagged posts. But if you want to report a guidelines violation for AI-assisted writing to the mods, just email the mods (contact link in the footer) subject “AI-assisted writing flag” or similar with a link to the post/comment. It works, I know, I’ve done it before. It takes maybe 60 seconds and there is no other way on the site (seemingly by OG design!) to guarantee human review but that email. |
| |
| ▲ | zahlman 8 hours ago | parent | next [-] | | > It works, I know, I’ve done it before. It takes maybe 60 seconds and there is no other way on the site (seemingly by OG design!) to guarantee human review but that email. It's a ton of friction compared to ordinary use of a forum; and while I've emailed several times myself, it comes with a sense of guilt (and a feeling that my "several" is probably approximately "several" above average). | | |
| ▲ | altairprime 8 hours ago | parent [-] | | Valid. It’s a big drawback of HN. I find it helps to report a perceived guidelines violation in “seems like” language rather than “is”, without demanding a specific mod outcome, in cases where I’m uncertain. That is noticeably distinct from “this is completely unacceptable” which I’ve said in a couple of instances, though I still tend to let the mods pick the outcome since that’s their job and I make a specific effort not to participate in sentencing decisions if at all possible. ps. I acknowledge as well that I’m exempt from feeling guilt for brain reasons, and so if it sounds like I’m not honoring what I would describe as a ‘completely normal’ human response, apologies; I’m trying my best given the lack of familiarity and intend no disrespect towards that reaction. |
| |
| ▲ | 152334H 4 hours ago | parent | prev [-] | | Never occurred to me to try that, because I assumed I would get banned for doing it, until today. | | |
| ▲ | altairprime 4 hours ago | parent [-] | | Nah, as long as you aren’t demanding and rude, you’ll either get a reply or not, and if you get a reply, it’ll either be “we’ll look into it”, “we looked into it and acted in some way”, or “we looked into it and decided it isn’t actionable”; often with some supporting explanation. (I suppose if you open with e.g. “wtf is wrong with you mods” they might well ask you to reconsider your approach or else clock a ban — I’ve never tried that!) |
|
|
|
| ▲ | postalcoder 9 hours ago | parent | prev [-] |
| I’ve actually been thinking about this exact idea for https://hcker.news/. Stay tuned, I’ve already started rolling out some comment filtering. |
| |
| ▲ | arrsingh 7 hours ago | parent [-] | | Oh I didnt know about this. Very cool. Is hcker.news only on web? Or is there a mobile app as well? | | |
|
