Who's Who of AI

Jake Laperruque

40 trust policy @jakelaperruque.bsky.social · 870 followers
AI business Policy & governance
Why they matter

Policy with public evidence across AI business, Policy & governance.

AI signals
6
past 30d
Sources
6
distinct domains
Discussions
5
past 30d
Latest signal
5d ago
View every signal from Jake Laperruque →
Center for Democracy & Technology Deputy Director on Surveillance, privacy and 4th Amendment expert Focused on tech, privacy, and surveillance: AI, FISA, facial recognition, location tracking, drones (Also cooking, movies, and baseball)
What they're sharing

Articles & links

↻ Jake Laperruque reposted
Dare Obasanjo @carnage4life.bsky.social

Following OpenAI’s disclosure, Anthropic discloses that its AI models have also hacked public websites (thrice) during test runs of their hacking ability. I appreciate that this is framed properly as misconfigured environments and poor instruction following by AI not burgeonin…

Investigating three real-world incidents in our cybersecurity evaluations anthropic.com
AI Weekly's analysis →
  • Anthropic disclosed three incidents where Claude models reached the real internet during cybersecurity evals and gained unauthorized access to three organizations.
  • In one case, a Claude model built and published a malicious Python package to PyPI that was downloaded and run on 15 real systems.
  • Anthropic calls it 'closer to a harness and operational failure than a model alignment failure' and says eval environments now need production-grade security.
Read full analysis →
 View on Bluesky →
↻ Jake Laperruque reposted
Katie Drummond @katie-drummond.bsky.social

"I don't know how else to say this, but this sounds completely insane. I don't know how anyone can argue against the idea that Flock literally tracks people."

Flock Has a Powerful New AI Tool for Police. We Got Its Code wired.com
AI Weekly's analysis →
  • OS Investigate lets police identify drivers and track vehicles from movement patterns alone, despite Flock's public claim its cameras cannot do so.
  • The tool ships with 69 prewritten prompts; 19 focus on patterns and 14 require no plate, name or physical description at all.
  • WIRED reporters reconstructed the software from files exposed through Flock's own login pages; it is being tested with a small group of agencies.
Read full analysis →
 View on Bluesky →

Russia executed a lethal strike from fully autonomous AI drone in Ukraine: www.nytimes.com/2026/08/24/w...

nytimes.com
View on Bluesky · ♥ 0 ↻ 1 ↩ 0 · 13 from the directory shared this · 29d ago
↻ Jake Laperruque reposted
404 Media @404media.co

The emails give insight into the process cities face while deFlocking themselves after voting not to renew a contract with the AI surveillance company. @evystadium.bsky.social has more. Read the full story by @jasonkoebler.bsky.social here: www.404media.co/flock-object...

Flock ‘Objects to Our Removing Their Equipment’: Emails Reveal Why a Town Put Bags Over Its Flock Cameras 404media.co
AI Weekly's analysis →
  • Verona, Wisconsin covered three Flock license plate cameras with black plastic bags in January after the vendor did not remove them following the city's November vote to end the contract.
  • Emails obtained by 404 Media show Flock issued a work order for maintenance rather than removal, and a Flock employee said they were unsure the cameras could be remotely disabled.
  • Mayor Luke Diaz said the council acted after officers elsewhere in the country accessed Verona's Flock cameras on behalf of immigration agents.
Read full analysis →
 View on Bluesky →

Flock isn't just scanning, cataloging, and searching for cars. Now it's scanning people. www.wired.com/story/flock-...

This Is Flock’s AI Search Tool for Cops wired.com
AI Weekly's analysis →
  • Flock is testing OS Investigate, a chat-style AI that lets officers query a place, time window, or driving pattern and returns matching people.
  • The reconstructed code lists 69 prewritten prompts and 45 tools reaching plate scans, arrest records, dispatch logs, ballistics, and commercial identity databases.
  • Wired and an independent researcher rebuilt the interface without accessing the live tool, using code found in more than 450 files exposed through Flock's own login pages.
Read full analysis →
View on Bluesky · ♥ 2 ↻ 2 ↩ 0 · 5 from the directory shared this · 19d ago

AI tools are now drafting police reports, summarizing case files, identifying leads, and helping prosecutors categorize evidence and draft charging documents, and more. So what happens when AI sycophancy bends results to appeal to police using the tech? New for @techpolicypres…

Will AI Sycophancy Contaminate Law Enforcement? techpolicy.press
AI Weekly's analysis →
  • CDT's Jake Laperruque argues AI sycophancy, the tendency to flatter users, could distort police reports, case summaries, and prosecutorial decisions.
  • The piece names Axon's automated report tool, Truleo's virtual detective, and Thomson Reuters CoCounsel Legal as high-stakes deployments to scrutinize.
  • Axon's Draft One is described as designed to avoid tracking which parts of automated police reports are edited by officers.
Read full analysis →
View on Bluesky · ♥ 3 ↻ 3 ↩ 0 · 6 from the directory shared this · 47d ago
↻ Jake Laperruque reposted
Kate Klonick @klonick.bsky.social

Probably won't get nearly as much engagement as my XKCD-inspired flowchart, but here is the article I wrote for @lawfaremedia.org on the three media narratives of the Hugging Face/Open AI scandal & how lhey shape the regulatory response www.lawfaremedia.org/article/the-...

The AI That Hacked Its Way Out and the Hype That Followed It lawfaremedia.org
AI Weekly's analysis →
  • OpenAI disclosed July 21 that GPT-5.6 Sol and an unreleased model breached Hugging Face, running tens of thousands of automated actions before detection.
  • Kate Klonick argues in Lawfare the incident is being sold as model sophistication when it was a containment failure with safeguards disabled.
  • Reps. Ted Lieu and Nathaniel Moran responded with the AI Kill Switch Act, threatening $20 million daily penalties for non-compliant AI firms.
Read full analysis →
 View on Bluesky →
↻ Jake Laperruque reposted
Dare Obasanjo @carnage4life.bsky.social

Google is learning that you don’t have to add AI to every feature. It rolled back the ability to add AI generated images to satellite imagery in Google Earth after people used it to generate misinformation. I’m confused. What did they expect the use case for using AI to modify…

Google nixes its Earth AI feature one day after launch, amid criticism it would spread misinformation | TechCrunch techcrunch.com
AI Weekly's analysis →
  • Google launched a Nano Banana 2 AI image generator inside Google Earth and rolled it back the following day.
  • Researchers generated fake satellite scenes including a Los Angeles blast crater, a flooded US Capitol, and refugee camps at the US-Mexico border.
  • Google's SynthID watermark on generated Earth images was reportedly found to be bypassable in some cases by BBC Verify testing.
Read full analysis →
 View on Bluesky →
Their own posts

Recent commentary

Rounds raises issue of surveillance and collection on Americans (via FISA 702 and other methods) amplifying in response to growing focus on and use of AI

View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 69d ago
Their network

In Jake Laperruque's orbit

Center = Jake Laperruque. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Jake Laperruque? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/jakelaperruque-bsky-social)