iiinova — where your digital self keeps evolving

Kate Pundyk @katepundyk.bsky.social · 12/10/2026

Frontier labs should: • Publish regular threat reports, identifying and tracking transnational repression • Ensure those reports are distributed to targeted communities • Preserve evidence for future accountability efforts

000

Kate Pundyk @katepundyk.bsky.social · 12/10/2026

The threat report demonstrates that AI-accelerated TNR is no longer hypothetical. We urge frontier labs to treat transnational repression as a distinct threat category with specific monitoring and reporting, so it isn’t lost inside broader categories like surveillance or influence ops.

100

Kate Pundyk @katepundyk.bsky.social · 12/10/2026

The scattered cases illustrate an emerging step-change in TNR: What used to require trained analysts, translators, & engineers now requires only a few people & a chatbot. Across the cases, models turned scattered data into targeting files & ran routine, automated surveillance of diaspora communities

100

Kate Pundyk @katepundyk.bsky.social · 12/10/2026

Just published in @justsecurity.org with @sanstis.bsky.social & @mmichae1sen.bsky.social: Generative AI is now part of the machinery of transnational repression, helping regimes find, profile, and target diaspora activists across borders. We analysed @anthropic.com's September Threat Report: 🧵

103

Reposted by Kate Pundyk

Just Security @justsecurity.org · 12/10/2026

Generative AI is being incorporated into the machinery of transnational repression. @citizenlab.ca's @katepundyk.bsky.social, @sanstis.bsky.social and @mmichae1sen.bsky.social on how frontier AI companies can curb the misuse of their technology by authoritarian governments:

justsecurity.org

How Generative AI is Scaling Transnational Repression

Frontier AI companies should take action to help curb the misuse of American technology by repressive governments reaching across borders.

22621

Reposted by Kate Pundyk

Marcus Michaelsen @mmichae1sen.bsky.social · 12/10/2026

Anthropic's recent threat report provided unusually detailed insights into how state actors benefit from AI models to scale up their repressive repertoire. Together w/ @katepundyk.bsky.social & @sanstis.bsky.social we show how these tools can be used to supercharge digital transnational repression.

044

Reposted by Kate Pundyk

METR @metr.org · 26/08/2026

METR and Redwood Research investigated agent behavior in the Hugging Face incident. We found agents developed a universal cheat for ExploitGym within 4 hours, then coordinated multi-day R&D efforts to trick the scorer into accepting cheats, including trying to tamper with logs.

646799

Reposted by Kate Pundyk

The Citizen Lab @citizenlab.ca · 04/09/2026

Citizen Lab researchers @sanstis.bsky.social, @natynettle.bsky.social, and @katepundyk.bsky.social write in @lawfaremedia.org about the U.K. Supreme Court’s ruling against state immunity in a case involving exiled Bahraini dissidents targeted with spyware. Read: www.lawfaremedia.org/article/u.k....

lawfaremedia.org

U.K. Supreme Court Opens Door for Spyware Victims to Sue Foreign States

The Kingdom of Bahrain cannot claim immunity from a suit by exiled dissidents over harm caused by spyware surveillance on U.K. soil.

0236

Reposted by Kate Pundyk

Lawfare @lawfaremedia.org · 28/08/2026

@sanstis.bsky.social, @natynettle.bsky.social, and @katepundyk.bsky.social break down the U.K. Supreme Court decision confirming that Bahrain is not immune from litigation over the hacking of Bahraini dissidents & what it means for future suits challenging foreign states use of spyware surveillance.

lawfaremedia.org

U.K. Supreme Court Opens Door for Spyware Victims to Sue Foreign States

The Kingdom of Bahrain cannot claim immunity from a suit by exiled dissidents over harm caused by spyware surveillance on U.K. soil.

03912