AI text detectors have been notoriously unreliable, but that's starting to change.
This year, Pangram keeps coming up as the trusted source in identifying AI-written text.
We sit down with Pangram CEO Max Spero to find out how the system was made, how much we should trust it, and where the line is between useful AI and AI slop.
0:00 Intro
00:21 Are AI Detectors Reliable?
01:27 90 Seconds on The Verge
03:15 Meet Pangram CEO
03:25 How Pangram Built Trust
05:51 Pangram’s Active Learning
08:30 Assessing Text and Interpretability
11:16 Why Perplexity Detectors Break
14:19 How to Read Pangram Results
20:03 Commonwealth Prize Controversy
23:30 Who Uses Pangram?
25:43 AI in Student Writing
28:14 Keeping Up With New Models
32:18 Beating Humanizers
34:11 Pangram’s Sustainability and Future
35:39 Slop Janitor Philosophy
Subscribe: http://goo.gl/G5RXGs
Like The Verge on Facebook: https://goo.gl/2P1aGc
Follow on Twitter: https://goo.gl/XTWX61
Follow on Instagram: https://goo.gl/7ZeLv
Watch The Vergecast on YouTube: https://bit.ly/40RFRkg
The Vergecast Podcast: https://bit.ly/3WQDexZ
Decoder with Nilay Patel: http://apple.co/3v29nDc
More about our podcasts: https://www.theverge.com/podcasts
Read More: http://www.theverge.com
Community guidelines: http://bit.ly/2D0hlAv
Wallpapers from The Verge: https://bit.ly/2xQXYJr
Shop our Verge merch store here: https://bit.ly/4kPCmEc
Subscribe to The Verge: https://bit.ly/3FT6n5S
Subscribe to The Vergecast on YouTube, new episodes on Tuesday and Friday: https://bit.ly/3I6nJtz
If you buy something from a Verge link, Vox Media may receive a commission without exerting any influence on editorial content.
For more information about our ethics policy, visit: https://bit.ly/3ZWTlLs