A.I. and Cultural Heritage Research: Risks and the Sins of Omission

The UCLA/Getty Program in the Conservation of Cultural Heritage and UCLA DataX invite you to attend a UCLA/Getty Program Distinguished Speaker Series lecture. On Tuesday, May 20 at 6:00 pm at the La Kretz Garden Pavilion, Damon Crockett (Preservation of Cultural Heritage, Yale) will deliver a talk titled A.I. and Cultural Heritage Research: Risks and the Sins of Omission. To register, please click here.

About the Presentation:
The sudden appearance and rapid proliferation of large language models (LLMs) in the past several years has brought about a reckoning over the use of intelligent machines in tasks previously done by humans. Cultural institutions, which bear a special responsibility to characterize and promote human excellence and ingenuity, are justifiably apprehensive about delegating important judgments to intelligent machines. In this talk, Crockett will argue that modern A.I. represents an increasingly faithful expression of human intelligence, and although there are associated risks, the risk profile is not what it may seem. The current generation of A.I. models are considerably less risky than their predecessors, because they are more intelligent, far easier to steer and none of their power is autonomous: it is derived entirely from contexts of deployment and is fully controllable by humans. Moreover, LLMs provide fresh opportunities to push toward the hypothetical end goal of cultural heritage research – an inclusive and accurate portrayal of the staggering diversity of human cultural expression.

About the Speaker:
Damon Crockett is lead scientist in the Lens Media Lab within the Institute for the Preservation of Cultural Heritage at Yale University. Previously, he was a postdoctoral associate in Yale’s Digital Humanities Lab, and has held appointments at the University of Chicago, UCLA and UC San Diego, where in 2015 he completed a Ph.D. in philosophy and cognitive science. He has spent the past 10 years applying A.I. to the study of culture, specializing in computer vision, data visualization, unsupervised learning and interpretability. More recently, his research attention has shifted to the “linguistic turn” in AI and the ways that large language models can be used productively in cultural heritage research.

This event is made possible by the generous sponsorship of Jeffrey P. Cunard.