“I used a detector which got the Pope in the Puffer jacket right. But then when I put in the Easter Bunny image that I made for my kids, it said that it was human-generated.”

TED Tech · Democracy in the Age of AI (2/5): When AI can fake reality, who can you trust? | Sam Gregory

Play from 9:53
9:53

Audio streams from the publisher’s own feed. Verbatim does not host or copy it.

What was described

tested a deepfake detector that correctly identified a Pope image but mis-labelled an Easter Bunny image as human-generated

Said nearby on the same subject

“We found that a vehicle going past typically has something like 28 images taken, but that can stretch up to 100 images”
♫
Said on Short Wave
Why it worked 28 images typical, up to 100 images Share this moment
“Any imaging you take, if there’s something on there you have to be able to identify it and refer out when needed.”
♫
Said on Dental Digest Podcast with Dr. Melissa Seibert
How they work Share this moment
“I have very simple text, very little text, but you have one image or one visual element ... intentional because it goes right back to driving down the highway, looking at a billboard.”
FD
Frederick Dunn
YouTube Creators Hub ·
Why it worked Share this moment
“If you test it before the 15th, build one brand kit… make 25 images across three formats, score every one, and count your correction time against what you do now.”
LR
Lonnie Rodriguez
AI Market Intel ·
How they work 25 images, three formats Share this moment

Moments like this, every Monday

The strongest new moments on your topics, in the speakers' own words. Free for three topics.

No card. No password. One-click unsubscribe.

Track a competitor?

Pro watches up to 25 names or phrases and emails you the morning after they are said.

See Pro, $29 a month

Verbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.