Flutterby™! : archival collections & AI

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

archival collections & AI

2025-07-21 17:55:52.633471+02 by Dan Lyke 0 comments

Springer Nature: AI & Society: How can we improve the diversity of archival collections with AI? Opportunities, risks, and solutions includes this quote from Javier de la Rosa of the National Library of Norway:

There are communities that have been mistreated in the history of the country, so when we trained these models, we fed them everything that was in the newspapers of the last century. We have encoded all that hate speech into the model. If you look for things like ugly people, some images might pop up that are very offensive to specific communities… We do have a huge responsibility when we release these models. These are not neutral artefacts in any way, they are biased machines. They are designed to exploit the patterns in the text and we really need to be careful with that.

Via Jessamyn

[ related topics: Language Interactive Drama Books Nature and environment Political Correctness Community Cryptography Artificial Intelligence Archival ]

comments in ascending chronological order (reverse):