this post was submitted on 27 Sep 2023
114 points (98.3% liked)
Firefox
20340 readers
93 users here now
/c/firefox
A place to discuss the news and latest developments on the open-source browser Firefox.
Rules
1. Adhere to the instance rules
2. Be kind to one another
3. Communicate in a civil manner
Reporting
If you would like to bring an issue to the moderators attention, please use the "Create Report" feature on the offending comment or post and it will be reviewed as time allows.
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Unfortunately, even OpenAI themselves took down their AI detection tool because it was too inaccurate. It's really, REALLY hard to detect AI writing with current technology, so any such addon would probably need to use a master list of articles that are manually flagged by human.
If you could detect AI authored stuff, couldn't you use that to train your LLM?
Suspect it would operate more on the basis of a person confirming that the article is of reasonable quality & accuracy.
So not unlike editors selecting what to publish, what to reject & what to send back for improvements.
If good articles by AI get accepted & poor articles by people get rejected, there may still be impacts, but at face value it might be sufficient for us seeking to read stuff.
It could be used to create a reward model like what is done right now with RLHF.
That said, it should actually be possible to make a bullshit detector that detects bullshit writing.