Go back
Writing

Substack now shows whether posts are 'AI-written' or not. The problem is the false flags and whether 'AI-written' can be accurately detected.

Substack's new AI-detection tool has some critics worried about false flags
Straight Arrow

Substack's new AI-detection tool has some critics worried about false flags

Substack is partnering with Pangram to let readers check how much of a post was AI-written. But even the company admits the tool isn't perfect.

Open Link

OddsRabbitSign up for OddsRabbit

Join communities like /c/Writing and never miss posts like this. Every signup donates a meal for a child in need. Learn more

Get the App About OddsRabbit
Goten_Arrigoni JingJunMa View comments 3 Post impressions 609
Stickers
Loading stickers...

Comments

JingJunMa's avatar
JingJunMa AmbassadorOP 2mo

Seeing a lot of controversy over this the past few days

3 Reply
Goten_Arrigoni's avatar
Goten_Arrigoni Ambassador 2mo

I've got false flagged when I was doing the TEFL certificate so I can understand the critiques lol.
I guess false positives shouldn't be an issue as long as it can be checked by a human. But it will definitely become an issue if false positives aren't submitted to human moderation!

2 Reply
JingJunMa's avatar
JingJunMa AmbassadorOP 2mo

I think the concern is that no one wants to read anything written by AI, and false positives really impact the users affected.

I don't think AI text detection can ever be properly done to be honest, how can text be scanned to check whether it's potentially AI or not?

1 Reply