BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Algorithm-mediated social learning - William Brady (Northwestern U
 niversity)
DTSTART:20260225T150000Z
DTEND:20260225T160000Z
UID:TALK240256@talks.cam.ac.uk
CONTACT:Yara Kyrychenko
DESCRIPTION:Social information is the currency of human culture\, and it i
 s increasingly encountered in online social networks where information flo
 w is controlled by algorithms. In practice\, this has generated a signific
 ant problem: social information encountered online is often highly unrepre
 sentative of all the people in a given social network. Focusing on the con
 text of politics\, I argue that the problem of unrepresentativeness is bes
 t explained as an emergent phenomenon that arises when humans and social m
 edia algorithms learn from one another (‘algorithm-mediated social learn
 ing’). I will present a series of experiments in the lab and in a large 
 field experiment that find social media algorithms exploit human social le
 arning biases toward PRestigious\, Ingroup\, Moral and Emotional informati
 on\, or PRIME information) in ways that distort our understanding of socia
 l norms. This work also examines consequences of algorithm-mediated social
  learning for intergroup relations\, and test newly designed algorithms th
 at mitigate unrepresentativeness\, offering possible paths toward healthie
 r online information environments.
LOCATION:Online
END:VEVENT
END:VCALENDAR