Source The New York Times · Published · Open the original ↗
Capability
I Tested a Popular A.I. Slop Detector. It Felt Empowering
Pangram excels at distinguishing chatbot-generated words from human writing. But it’s not reliable for spotting artificial images.
Recently, as I was scrolling through my phone, I realized I was suffering from a new kind of fatigue: slop-induced burnout. Almost all the posts I was reading on social media were riddled with bullet points, emojis and em dashes. They clearly looked like the work of a chatbot .
This month, my suspicions were confirmed by a web tool that is getting buzz: Pangram , which uses artificial intelligence to detect when words were generated with chatbots. Users can simply paste in a block of text for Pangram to scan, and the tool calculates a percentage indicating how much of the text was churned out by a bot or written by a person.
The opening of the report, quoted unchanged from The New York Times. The full text continues at the source.