How many languages will AI speak when it wipes out humanity?

Date and time: 25/04/24, 5pm CET


Eneko Agirre, Ixa Research Group, Hitz Basque Centre for Language Tecnhology


Generative AI has advanced dramatically. In particular, Large Language Models (LLMs) such as GPT have raised expectations of superhuman intelligence and catastrophic predictions, overshadowing more subtle and immediate benefits and problems. For example, GLMs are especially effective in the most widely spoken languages, but what about the rest of the languages? In this talk we will review how MLGs are built, their risks, and, above all, the possible strategies for MLGs to offer the same quality in languages with fewer resources, taking Basque as a prominent example. GB.


Graduate in Computer Science (1990, UPV/EHU, Special Prize), Master of Science in Knowledge Based Systems (1991, Edinburgh University), and Doctor in Computer Science (1999, UPV/EHU). Professor since 1995, now full professor (contract professor). Director of the LSI department (2017-2019). National Research Award in Computing 2021. ACL fellow since 2021.

He directs the HiTZ center ( and the IXA Group (, group A for the GV. He has supervised 15 theses. He has been a researcher or visiting professor at New Mexico State, Melbourne, Southern California, Stanford and New York universities. He has participated in more than 50 research projects, 15 of them as principal investigator, in competitive calls in Europe, Spain and the Basque Country. He has had 9 transfer contracts with companies. He has more than 200 articles, including 13 top-level journals (Q1 or Q2) and 49 top-level conferences (SCIE class 1/2). He has 5 six-year terms. He has been or is a member of the editorial board of the two leading journals in his field (Computational Linguistics, TACL) and the journal JAIR, as well as president of SIGLEX. Founder of the STARSEM (*SEM) conference, now in its 11th edition. He has won three best paper awards and three Google research awards. He has given more than 20 invited lectures, most of them international. The research topics are artificial intelligence and language technology based on deep learning.

Mandatory registration via Zoom:

Enlaces relacionados