THIS TALK WILL BE RESCHEDULED. New date TBD, thanks for your understanding!

Join the Libraries for a talk by Aliya Bhatia and Gabriel Nicholas from the Center for Democracy & Technology on their recent report, "Lost in Translation: Large Language Models in Non-English Content Analysis"

This will be a hybrid event, join the Zoom here.

Today, many social media companies build their automated content moderation systems using large language models. Some companies, including Facebook, Google, and Bumble, use these models to moderate content at scale in dozens, or even hundreds of languages. Yet, many of these models are trained on far more English-language training data than in any other language due to the fact that many languages, spoken predominantly in the Global South, have fewer training datasets available. This talk presents evidence that language models may not be up to the very language- and culture-dependent task of content moderation, explores how these shortcomings may threaten global free speech online, and offers suggestions on how to improve language equity in AI writ large.

Please click the link below to join the webinar:
https://haverford.zoom.us/j/99791693138?pwd=c09Ec3V1K3MvYVdvQnBJSFpNQmdJUT09
Passcode: 145067
Or One tap mobile :
    +16469313860,,99791693138#,,,,*145067# US
    +19292056099,,99791693138#,,,,*145067# US (New York)
Or Telephone:
    Dial(for higher quality, dial a number based on your current location):
    +1 646 931 3860 US
    +1 929 205 6099 US (New York)
    +1 301 715 8592 US (Washington DC)
    +1 305 224 1968 US
    +1 309 205 3325 US
    +1 312 626 6799 US (Chicago)
    +1 564 217 2000 US
    +1 669 444 9171 US
    +1 669 900 6833 US (San Jose)
    +1 689 278 1000 US
    +1 719 359 4580 US
    +1 253 205 0468 US
    +1 253 215 8782 US (Tacoma)
    +1 346 248 7799 US (Houston)
    +1 360 209 5623 US
    +1 386 347 5053 US
    +1 507 473 4847 US
Webinar ID: 997 9169 3138
Passcode: 145067
    International numbers available: https://haverford.zoom.us/u/abwtZL8uct