Sign in

Lisa LeVasseur

@lalevasseur.bsky.social
236 followers 929 following 98 posts

Founder of Inter Safety Labs, measuring invisible tech harms.

PostsRepliesMedia
Reposted by Lisa LeVasseur
Alex de Campi @alexdecampi.bsky.social · 15/10/2026
UNMUTE this is GLORIOUS
7950321 0
Lisa LeVasseur @lalevasseur.bsky.social · 08/10/2026
My 1st job was SW Engr in the 80s, z8000 assembly language. Ironically sw dev is more about assembling [3rd party pts] today. been interesting to watch the decades long decline in the rigor of the craft, esp sw architecture. We can’t possibly be surprised that sw agents are exploiting sloppy code.
010 0
Lisa LeVasseur @lalevasseur.bsky.social · 06/10/2026
ugh, this timeline with its insane trifecta of “labor is expendable”, “consumers are expendable” and “ natural resources are/the planet is expendable”. I mean, think it through, brainiac zillionaires. 🤨
000 0
Reposted by Lisa LeVasseur
Kat Abughazaleh @katmabu.bsky.social · 29/09/2026
I literally don’t know a single woman who hasn’t been raped, stalked, groped, harassed, or sexually violated in some way at some point in their life. If you choose to doubt and dox survivors, just know every woman in your life is disgusted by you.
768286765983 0
Reposted by Lisa LeVasseur
Zach Edwards @thezedwards.bsky.social · 27/09/2026
But none of the new ad tech standards require sharing these details publicly - ad tech won't do this by default unless they are required by law. But if one enterprising legislator out there was able to push through an effort to require the ad tech industry to publicly share new data after auctions..
122 0
Reposted by Lisa LeVasseur
Zach Edwards @thezedwards.bsky.social · 27/09/2026
There are countless ad tech companies who play fast and loose with what it means to be a data broker so they don't register as a data broker, but they still collect profiles on people and help 3rd parties target ads to those ad segments. But those folks get to largely hide and never face scrutiny.
141 0
Reposted by Lisa LeVasseur
Dr. Damien P. Williams, dread portent down from a mountain cave @wolven.blacksky.app · 19/09/2026
Again: "Tilly Norwood" is not a person; it is a shoddy, and also pretty gross puppet. Stop referring to it as "she," and start asking whether there might be something a bit off about wanting to promote a sexualized but also childlike "actor" that can never tell its creators or users "no."
452961739 0
Lisa LeVasseur @lalevasseur.bsky.social · 17/09/2026
Btw this is the real problem with throwing automated “AI“ acceleration onto a dumpster of imperfect extant software.
000 0
Reposted by Lisa LeVasseur
Zach Edwards @thezedwards.bsky.social · 29/08/2026
We launched decryptads.com two weeks ago to help people dumpster dive on the programmatic ad tech ecosystem – the ad tech companies, publishers, apps and accountIDs. We’re building my dream ad tech research tool, but I want to explain why it kicks ass. 🧵
decryptads.com
DecryptAds — From Hidden Flows to Public Insight
DecryptAds maps programmatic ad-tech supply chains from public files — ads.txt, sellers.json, and related signals — so you can trace hidden flows and surface invalid supply.
2125 0
Lisa LeVasseur @lalevasseur.bsky.social · 27/08/2026
Well jeez. Two icons whose existence made a lasting impression on me gone in as many days. Grateful for the seeds of joy they watered.
010 0
Lisa LeVasseur @lalevasseur.bsky.social · 16/08/2026
Ugh would like to never see “machine expression” again. “Machine extrusion” or machine-extruded text would be better.
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 16/08/2026
Such tricky work here. heaven forbid we further fortify corporate (or worse: product) rights over human rights. Anyone else find themselves fantasizing about Ted Nelson’s original architecture of hypertext where there’s only a single copy?
010 0
Reposted by Lisa LeVasseur
SASSI South @sassisouth.bsky.social · 15/08/2026
In addition to Flock and Axon, another company that we should oppose is Peregrine aka Palantir-lite. It was started by a former Palantir executive and sells a predictive policing tool to local policing agencies. They are funded by all of the technofascist VC companies. pulse2.com/peregrine-te...
pulse2.com
Peregrine Technologies Raises $250 Million Series D At $6.8 Billion Valuation
Peregrine Technologies announced that it has raised $250 million in Series D financing at a $6.8 billion valuation. The round was led by existing investors, including Fifth Down Capital, Sequoia Capit...
152 0
Reposted by Lisa LeVasseur
Meredith Whittaker @meredithmeredith.bsky.social · 02/12/2024
📢NEW: 'Open' AI systems aren't open. The vague term, combined w frothy AI hype is (mis)shaping policy & practice, assuming 'open source' AI democratizes access & addresses power concentration. It doesn't. @smw.bsky.social, @davidthewid.bsky.social & I correct the record👇 nature.com/articles/s41...
nature.com
Why ‘open’ AI systems are actually closed, and why this matters - Nature
A review of the literature on artificial intelligence systems to examine openness reveals that open AI systems are actually closed, as they are highly dependent on the resources of a few large corpora...
271027346 0
Lisa LeVasseur @lalevasseur.bsky.social · 13/08/2026
I feel like we should be talking more about the demise of paragraphs in synthetically produced text. I find it more readable. Yet worry that it’s an indicator of further decline of my ability to read long text and process complex ideas. You?
100 0
Reposted by Lisa LeVasseur
Evan Greer @evangreer.bsky.social · 11/08/2026
I know everything feels terrible and doomy but I've been working on surveillance issues for most of my adult life and I have never seen so much energy, momentum, and concrete VICTORIES happening as what we are seeing right now on Flock cameras and facial recognition in venues. Keep. Going.
3650471138 0
Reposted by Lisa LeVasseur
Dr. Damien P. Williams, dread portent down from a mountain cave @wolven.blacksky.app · 08/08/2026
Anyone running middle or high school classes where you give kids cheap blank phones & teach them to deconstruct them? I feel like getting people at an early age w/ things like "what do API calls do?" & "what are your phones/apps doing when you're not touching them?" would be good on multiple levels…
814933 0
Lisa LeVasseur @lalevasseur.bsky.social · 03/08/2026
Love this framing @tressiemcphd.bsky.social data politics doesn’t happen without data economics. Central to this must also be acknowledging the immorality of personal data as a market/good. www.nytimes.com/2026/08/03/o...
nytimes.com
Opinion | A Reading List for Our Age of A.I. (Gift Article)
A.I. is a political theory. And it is gobbling up our democracy.
14314 0
Reposted by Lisa LeVasseur
Mike Masnick @masnick.com · 03/08/2026
This got me to finally read the entire New Yorker article on the Tates and... yeah. I don't see how what they're doing is any different than what everyone has accused Jeffrey Epstein of doing. Except they're *more public* about admitting to what they're doing AND getting young men to join in.
201633448 0
Reposted by Lisa LeVasseur
Thomas Spooks 🎃👻💀 @thomasfuchs.at · 31/07/2026
The anthropomorphization of "AI" is literally PR speak handed down by "AI" companies that want you to think it's magic and also to think that they're not responsible for problems.
110630 0
Reposted by Lisa LeVasseur
Karl Bode @karlbode.com · 31/07/2026
most tech outlets spent the week burying the lede that this was a completely avoidable error (or intentional marketing decision) by a human being, and not a malevolent self aware cyber consciousness that slipped its leash
wired.com
OpenAI’s Hacking Debacle Comes Down to Human Error
If the generative AI giant had followed well-known security best practices, it’s likely that its AI agent would never have escaped to the open internet and hacked multiple companies.
818968 0
Lisa LeVasseur @lalevasseur.bsky.social · 27/07/2026
Corruption and deceit it is.
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 27/07/2026
110 0
Reposted by Lisa LeVasseur
Tressie McMillan Cottom @tressiemcphd.bsky.social · 26/07/2026
As good as many of you said. I really appreciate the unambiguous point that this primarily a racist moral panic before unpacking the nuanced second-order explanations. dannykatch.substack.com/p/we-used-to...
dannykatch.substack.com
We used to be able to buy things in this country
Locking up toothpaste was a warning sign that our executive class was crazy
213027713 0
Lisa LeVasseur @lalevasseur.bsky.social · 23/07/2026
A good thread. Data brokers are a symptom though. The disease is regarding personal info as a good/product. It’s a morally bankrupt pov that society must reject.
020 0
Lisa LeVasseur @lalevasseur.bsky.social · 23/07/2026
1000% this.
010 0
Reposted by Lisa LeVasseur
By Cory Doctorow (GPG 0xBF3D9110957E5F4C) @doctorow.pluralistic.net · 23/07/2026
It's especially instructive to compare this process to the steps you have to take in order to "opt in" to having a data broker open a file on you and stuff it full of your sensitive, personal information, which is then sold to all comers: * Step one: Exist. * Step two: There is no step two. 40/
1214 0
Reposted by Lisa LeVasseur
Dr Heidy Khlaaf (هايدي خلاف) @heidykhlaaf.bsky.social · 22/07/2026
The coverage on this OpenAI incident is abysmal. Use of the terms "rogue" and "loss of human control" lead to groupthink as people lack the critical skills to understand the difference between "autonomy" and faulty reward functions in AI on a task it was directed and given access to do.
217870 0
Reposted by Lisa LeVasseur
Molly Knight @mollyknight.bsky.social · 21/07/2026
Oh! They gave millions to Trump lol. people.com/taylor-farms...
people.com
Taylor Farms Met with Trump White House Days Before FDA Backtrack on Cyclosporiasis Findings, Sparking Controversy
Executives at lettuce supplier Taylor Farms met with Trump White House officials days before the FDA walked back a ‘false positive’ detection of the cyclosporiasis parasite that is causing ‘explosive’...
30856280 0
Reposted by Lisa LeVasseur
Jon P @jdp23.thenexus.today · 21/07/2026
In terms of "product safety", ISL's site has some excellent framing here, that puts the focus squarely where it should be: "We believe everyone deserves to know a product’s safety risks."
001 0
Lisa LeVasseur @lalevasseur.bsky.social · 21/07/2026
💯
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 21/07/2026
It just occurred to me what my issue is with “safety” framing: it implies personal responsibility. “Product safety” prioritizes manufacturer responsibility.
111 0
Reposted by Lisa LeVasseur
Emily M. Bender @emilymbender.bsky.social · 19/07/2026
How about a jet-lagged live posting while I read? Trying to read this because I keep getting tagged in posts about it, but the first paragraph already contains synthetic text (a quote from Claude). I will skip that bit and persevere... www.theguardian.com/technology/2...
theguardian.com
Could AI be conscious?
Experts believe it’s at least possible. We urgently need a plan to navigate the ethical implications
24475168 0
Lisa LeVasseur @lalevasseur.bsky.social · 20/07/2026
So it’s either corruption and deceit or incompetence? The Sophie’s choice no one wants.
001 0
Reposted by Lisa LeVasseur
Shannon Vallor @shannonvallor.bsky.social · 20/07/2026
AI labs: “hey, what if we could take the Dunning-Kruger effect, bottle it in software, and sell it to every workplace, government office, and school on the planet? Now that’s innovation!”
23318 0
Lisa LeVasseur @lalevasseur.bsky.social · 18/07/2026
Could this be the world’s way of telling us that iceberg lettuce is more of a pathogenic carrier and less of an actual vegetable? (it doesn’t have any actual nutritional value on its most hygienic day, does it?)
000 0
Reposted by Lisa LeVasseur
Greg Pak @gregpak.net · 17/07/2026
This is exactly it -- if you make a kid use generative "A.i." or an LLM, you're telling them they're not good enough. That their own minds and ideas and experiences and abilities are not enough. That's a terrible, thoughtless, cruel thing to do. Shame on Kaepernick and everyone involved in Lumi.
5780213 0
Reposted by Lisa LeVasseur
Johnathan Perkins @johnathanperk.bsky.social · 07/07/2026
Hey so, if you use Waymo driverless car services, you are aiding and abetting fascism. Waymo helps ICE agents locate people to terrorize and abduct. The company willingly shares all facial recognition, license plate and GPS data with ICE and other law enforcement agencies—no subpoena required.
static.klipy.com
Waymo Self-Driving Jaguar at Traffic Light
ALT: Waymo Self-Driving Jaguar at Traffic Light
35164503437 0
Reposted by Lisa LeVasseur
Katie (Kathryn) Conrad @kconrad.bsky.social · 07/07/2026
In case folks - patients or medical professionals - might find it useful, here's an edited version of a letter I sent to my own physician(s) about the reasons I refuse the use of ambient scribes.
kconrad.substack.com
A letter to my doctor
On the refusal of ambient AI scribes
21612 0
Lisa LeVasseur @lalevasseur.bsky.social · 25/06/2026
A moment of joy afternoon pick me up: youtu.be/_o2RS8WfcbY?...
youtu.be
Bobby McFerrin & Chick Corea - Spain - LIVE HD
YouTube video by Zycopolis TV
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 24/06/2026
🤣
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 24/06/2026
may I suggest “commercial surveillance” v. “online surveillance”? @internetsafetylabs.bsky.social has done a boatload of research on data leakage by design in apps used by children. appmicroscope.org/app/1579/
appmicroscope.org
App Microscope
000 0
Reposted by Lisa LeVasseur
Aaron Stewart-Ahn @badideas.bsky.social · 20/06/2026
Cannot fucking believe an external 4TB SSD drive that cost me $200 last year is now $1100. Had to check receipts to make sure. They’re really out there trying to keep you from having a PC, phone, or a gaming console so they can have hallucinating, polluting chatbots that drop bombs on kids’ schools.
6535701105 0
Reposted by Lisa LeVasseur
Meredith Shiner @meredithshiner.com · 19/06/2026
"The FAA is working with Palantir to see if AI can help prevent collisions and close calls on airport runways" is a sentence that makes me never want to get on an airplane again.
361309359 0
Reposted by Lisa LeVasseur
Crystal Fleming is on sabbatical @alwaystheself.bsky.social · 17/06/2026
He kept coming back, repeatedly (in the way AI often loops) to this idea of AI (specifically Claude) representing "a position without an address", and it became increasingly clear to me that there was some odd mystification happening. Of course LLM products have an address. It's a corporate address.
202698245 0
Reposted by Lisa LeVasseur
Emily M. Bender @emilymbender.bsky.social · 16/06/2026
Second, what LLMs do is make papier-mâché of their input. So you could put in the best, most carefully curated sources in the world and what would come out would still be some kind of slurry. >>
213330 0
Lisa LeVasseur @lalevasseur.bsky.social · 16/06/2026
@melanie-ensign.bsky.social and I had an illuminating chat a few weeks ago about my Product Safety Resistance Playbook. Her words convinced me that I *am* being too nice to manufacturers. Wrote about it here: product-safety-for-the-information-age.ghost.io/melanie-says...
000 0
Lisa LeVasseur @lalevasseur.bsky.social · 12/06/2026
Abolition is the only sane response.
000 0
Reposted by Lisa LeVasseur
Paris Marx @parismarx.com · 10/06/2026
more countries need to follow this example
1530151 0