Greg Leppert weighs in on AI developers' use of written texts, rather than simply digital ones, to train LLMs.

Bruce Schneier and Nathan Sanders speak to fears about AI replacing humans' jobs, opining that this will only happen when it outperforms us in terms of speed, scale, scope, and/or sophistication.

Faculty Associate Ben Green exposes some of the challenges to Elon Musk's purported goal of using AI to reduce government spending.

Jonathan Bellack argues that Bluesky is nearing the end of the timeframe in which it can build a truly participatory content moderation process.

Mailyn Fidler's latest research paper posits that internet fragmentation hasn't been as inward-looking as theorists have often suggested.

Affiliate Bruce Schneier testifies before the House Committee on Oversight and Government Reform concerning DOGE's access to and use of Americans' data, reminding congresspersons that "data is power." ...

Sahana Udupa explores the dark side of encrypted messaging apps, exploring how hateful messages (which she dubs "extreme speech") can flourish in such apps because smaller-scale communications make ...

Rumman Chowdhury cautions that, no matter how useful AI can be, users shouldn't have it do their thinking for them.

BKC Affiliate Kathy Pham introduces The United States Digital Service Origins, an oral history of the USDS featuring nearly 50 interviews telling the story of technology and government.

With Sacha Alanoca and Kevin Klyman, Affiliates Shira Gur-Arieh and Tom Zick “present a taxonomy to map the global landscape of AI regulation.” “[T]his taxonomy serves as a tool which can resist time ...

In The Atlantic, Faculty Director Jonathan Zittrain probes policy questions raised by the “assumptions” large language models make about users. “Gleaning a model’s assumptions is just the beginning.

Scott Brennen and BKC Affiliate Zeve Sanderson warn about Congressional attempts to undermine states' ability to close the AI regulatory gap. They focus in particular on attempts to instate a ...