Will language models help us climb the Tower of Babel?

Presenter: Karen Livescu
Time & Date: 7-9 PM
Monday October 19, 2026
Location: The Map Room
1949 N. Hoyne
(Free, Limited to first 50 Attendees)

Summary

Large language models, like ChatGPT and Claude, are getting increasingly powerful and pervasive. But how pervasive are they really? Do they work everywhere and for everyone? In order to work well, LLMs need to be trained on enormous amounts of data, mainly written and/or spoken language. We have those enormous amounts of data in English, and in a handful of other high-resource languages. But for most of the world’s 7,000 languages, and for many dialects within them, we do not have nearly enough data. How can we get the benefits of LLMs, and artificial intelligence more broadly, to all of the world’s inhabitants and not only to those of us in high-resource places and communities?

(Side note: There are also many dangers of AI, as we have all been hearing in the news and social media. That is not the main topic of this conversation, but I am happy to discuss how it all fits in.)

Download the event flyer (PDF, 1.2MB)