Jake Laperruque
banner
jakelaperruque.bsky.social
Jake Laperruque
@jakelaperruque.bsky.social
Center for Democracy & Technology Deputy Director on Surveillance. I talk about tech, privacy, and surveillance policy: AI, FISA, facial recognition, location tracking, reproductive health privacy (Also cats, movies, cooking and baseball)
This is a real pitch material from an AI vendor pushing its tool to police and prosecutors, recommending they take a “too long; didn’t read” approach to evidence and investigations.

Does this make you feel safe?
November 13, 2025 at 8:06 PM
And this Explainer on the convoluted legal landscape of the border search exception:
cdt.org/insights/no-...
November 13, 2025 at 5:56 PM
DHS does not need probable cause suspicion to take your phone when you enter the US and fully copy it into its databases. It can then scan through it for any reason, and use AI tools to effortlessly pull up sensitive info

It’s a massive & easily abused loophole for basic 4th Amendment protections
November 13, 2025 at 5:54 PM
Cellebrite’s AI device scanning and analytics tool also lets users build out contact maps across various communications apps, and create detailed maps of individuals movements based on data within a phone. In a tutorial Cellebrite says it lets police be a “fly on the wall” in your life
November 13, 2025 at 5:52 PM
What happens when AI tools are applied to this massive database? The prospects are alarming. For example, tools from Cellebrite (which DHS contracts with) allow users to rapidly scan through photo libraries for certain types of images, such as “Money,” “Tattoos,” “Nudity,” “Drugs,” “Screenshots”
November 13, 2025 at 5:50 PM
This response from Google really makes no sense to me.

"We only let government personnel use the app" yeah, the whole complaint is that your helping cops use a sketchy facial recogition app to detain people (while also blocking off basic police oversight/accountability tools)
November 13, 2025 at 4:47 PM
Axon wants police to push a button and let AI write police reports and create the formal written record of officer encounters and use of force incidents.

Does this make you feel safe?
November 7, 2025 at 10:01 PM
This bit from Ars Trchnica seems to be the most important takeaway: The trillion dollar possibility will get lots of attention, but it’s also based on bold long term goals. Musk can be a replacement level CEO and still reap obscene rewards.
arstechnica.com/tech-policy/...
November 6, 2025 at 10:21 PM
Closing slide of a tech policy presentation from literally two hours ago. The Butlarian Jihad art/merch must flow!
November 6, 2025 at 8:53 PM
November 6, 2025 at 8:12 PM
Smart takedown by @emanuelmaiberg.bsky.social on Meta claim that it's smart glasses are no different than a phone:
It's not just that recording with these glasses is (obviously) more inconspicuous, Meta's made central design and selling point that they're inconspicuous!

Have cake, eat it too.
November 5, 2025 at 4:51 PM
November 4, 2025 at 7:00 PM
Guillotin was lucky he didn't get guillotin-ed
November 3, 2025 at 5:55 PM
Treating facial recognition as a definitive ID is at odds with virtually every law enforcement agency in America that has policies & procedures for using the tech.
Don't treat a match as a defintive ID is perhaps the most basic rule of the tech, including by DHS. These actions violate agency rules:
October 29, 2025 at 5:00 PM
"Nearly half of ICE recruits failing open book written exam on constitutional and legal requirements for conducting stops, searches, and seizures:
www.nbcnews.com/politics/imm...
October 27, 2025 at 4:38 PM
😬
October 16, 2025 at 3:41 PM
An interesting (and frustrating) contrast: Even as California continues to be an ever-active national leader on consumer privacy, it's fallen behind other states on surveillance issues such as facial recognition:
www.techpolicy.press/status-of-st...
October 10, 2025 at 7:53 PM
Rules for warrantless border searches of electronic devices have become complex and convoluted - you can @cdt.org analysis if you want to see the full legal landscape of what rules apply at each airport across the country:
cdt.org/insights/no-...
October 9, 2025 at 6:26 PM
Section 230 is a liability shield for hosting third party content online - there's potentially an argument AI chatbots are covered, but I think it would be very tough, common view (even from Altman) is that the company's chatbot is speech from the company

www.semafor.com/article/08/2...
October 9, 2025 at 2:53 PM
"New lists! New lists!"
October 8, 2025 at 11:58 PM
COMMUNISM WAS JUST A RED HERRING!
October 3, 2025 at 11:41 PM
The full written rules for use of body cameras by DHS personnel is available here:
www.dhs.gov/sites/defaul...

Meta's glasses squarely within the policy definition of body cameras
September 24, 2025 at 1:31 PM
Is it surprising that the focus group designed to exclusively be dissatisfied Trump voters included lots of dissatisfied Trump voters?
September 17, 2025 at 6:05 PM
Couple months ago this guy did an interview containing one of the more wild and self-revealing things I've heard from the set of rich AI-utopian tech bros:
www.wired.com/story/big-in...
September 17, 2025 at 3:42 PM
Also, interestingly it seems like this set of rules isn't followed for some of the most famous instances of murder that didn't result in a conviction:
September 11, 2025 at 2:30 PM