Gordon Pennycook

Associate Professor of Psychology

Gordon Pennycook is an associate professor of psychology who studies human reasoning and decision making, including misinformation interventions, such as labeling and warnings.

Pennycook says:

“It is possible for voluntary disclosures to backfire if people who are being honest (by disclosing AI use) are punished for it, while people who are dishonest (by not disclosing AI use) avoid punishment. For this to be effective, it has to be possible to detect people who use AI but fail to disclose it. My sense of the current moment is that a lot of artists would place the risks of being discovered over the risks of getting caught for not disclosing – time will tell.”

Trystan Sterling Goetze

Director of the Sue and Harry E. Bovay Program

Trystan Goetze, assistant teaching professor, says there needs to be some way of verifying the labels for accountability.

Goetze says:

“Labelling AI-generated and human-generated content is one way to help consumers find what they want and avoid what they dislike. And many consumers want to eschew AI-generated media for various principled reasons. 

“Mandating these AI labels could help restore the trust of consumers in a ‘slop’-filled online landscape. But since Spotify isn’t requiring the labels, this will only go so far—it’s like a grocery store making it optional for products to be labelled with their ingredients. 

“There needs to be some way of verifying the labels for accountability, just like the FDA can investigate food producers. Something like the C2PA content credential system could help with verification, but the technology isn’t perfect, and forgeries are always a possibility.”