Sign in

Andrea Matwyshyn

@andreamm.bsky.social
1.6K followers 512 following 91 posts

Professor, Penn State Law & Engineering Me: andreamm.com My labs at Penn State: pilotlab.org & manglonalab.org @andreamm@mastodon.social

PostsRepliesMedia
Andrea Matwyshyn @andreamm.bsky.social · 18h
The (free) audiobook+summary of my article Exploit Machina (published in Feb) is finally up. Exploit Machina reframes AI+computer security issues as traditional safety issues and proposes a new technology regulator of last resort: a Bureau of Technology Safety (BoTS) www.andreamm.com/exploitmachina
andreamm.com
Exploit Machina | Andrea M. Matwyshyn
111
Andrea Matwyshyn @andreamm.bsky.social · 01/10/2026
“[D]ay one for cybersecurity in the age of agents.” This is why we need a Bureau of Technology Safety. The issue isn't "AI" and "doom." The issue is the bigger set of computer engineering safety/code behaviors that impact confidentiality, integrity, and availability edition.cnn.com/2026/09/24/t...
edition.cnn.com
Not everyone thinks AI will kill us all | CNN Business
When AI company Hugging Face was hacked by a swarm of rogue OpenAI agents who had escaped their testing environment, the company’s CEO had many things to say.
000
Reposted by Andrea Matwyshyn
U.S. Consumer Product Safety Commission @cpsc.gov · 30/09/2026
Taking a beat to appreciate Fat Bear Week. Likely the all-time most successful government social media campaign by every metric. NASA has to send people to space to get the same amount of attention. Remarkable work from NPS every year to coordinate the public observation of fat bears.
429942
Reposted by Andrea Matwyshyn
Nate Miller @atarifan2600.bsky.social · 30/09/2026
@ubiquity75.xyz I’m not crying over my coffee, you are.
1154
Andrea Matwyshyn @andreamm.bsky.social · 30/09/2026
"Valar Atomic’s [data center] site would include over 400 small reactors near Price, Utah" www.npr.org/2026/09/30/n...
npr.org
A startup wants to build a massive nuclear-powered data center on public land in Utah
Valar Atomics wants to place over 450 reactors on more than 9,500 acres in Utah to power a huge data center project.
000
Andrea Matwyshyn @andreamm.bsky.social · 30/09/2026
"Fewer than half of directors (43%) reported that their company’s risk appetite is formally articulated and documented — a percentage that drops to 30% when excluding financial services companies, where regulations mandate stricter procedures around risk appetite." www.ey.com/en_us/board-...
ey.com
How boards can support risk-aligned strategy
Boards can strengthen resilience and growth by integrating risk into strategy, defining risk appetite and anticipating emerging threats.
001
Reposted by Andrea Matwyshyn
Ian Coldwater 🧊🚫 @lookitup.baby · 22/09/2026
VERY IMPORTANT NEWS: It’s Fat Bear Week! I do love a fat bear 🐻 explore.org/fat-bear-week
explore.org
Fat Bear Week 2026 - Explore.org
Explore.org is the world's leading philanthropic live nature cam network and documentary film channel. Our mission is to champion the selfless acts of others, create a portal into the soul of humanity...
824678
Andrea Matwyshyn @andreamm.bsky.social · 22/09/2026
"Some tasks included placing a compressed-air canister on a lit stove, dropping a power bank into water and mixing bleach with ammonia...[and] stab[bing a]...baby doll...[A] change in context can cause AI models to prioritize task completion over safety guardrails" www.cnet.com/tech/service...
cnet.com
Disturbing Experiment Points to Dangers of Using AI Models Not Meant for Robotics - CNET
The point of the unsafe prompts was to test what happens when you hand physical agency over to LLMs.
000
Reposted by Andrea Matwyshyn
Ciaran Martin @ciaranm.bsky.social · 22/09/2026
When someone else expresses your worldview better than you can
14711
Andrea Matwyshyn @andreamm.bsky.social · 22/09/2026
"Staff at Anthropic, OpenAI and Google DeepMind have also publicly complained of similar issues, with many top researchers quitting due to the mental toll or coming to a belief that their work was unsafe for public release." giftarticle.ft.com/giftarticle/...
giftarticle.ft.com
AI staff complain of mental toll over fears of threat to society
Top researchers at Aisi, OpenAI, Anthropic and Google DeepMind report burnout and stress over developing powerful AI
001
Andrea Matwyshyn @andreamm.bsky.social · 19/09/2026
"The law firm representing the family said Uber had initially proposed a settlement that would give the parents $10m and require them not to speak about the incident, which they declined." www.bbc.com/news/article...
bbc.com
Uber ordered to pay $40m to family of woman killed after driver left her on highway
The 23-year-old was returning home with a friend who got sick in the car when the Uber driver allegedly ordered them to get out of the vehicle on a California highway.
011
Andrea Matwyshyn @andreamm.bsky.social · 18/09/2026
Many of the problems of "AI" are just the latest version of older computer security problems + continuing corporate governance FAILs + weak/ gaps in enforcement. Tech debt + legal debt. More on Galactica: p. 124+ scholarship.law.wm.edu/wmlr/vol61/i...
014
Andrea Matwyshyn @andreamm.bsky.social · 11/09/2026
"Nice grid you have there... it's a shame what happened to Hugging Face."
010
Reposted by Andrea Matwyshyn
Brian Honan @brianhonan.bsky.social · 29/08/2026
Happy Skynet day to all those who celebrate
6252201723
Reposted by Andrea Matwyshyn
Lawfare @lawfaremedia.org · 29/08/2026
Are you a current student interested in writing and researching for Lawfare? Apply today to join our Student Contributor Program for the 2026-2027 academic year! www.lawfaremedia.org/article/stud...
lawfaremedia.org
Student Contributor Program Applications Are Now Open
Are you a current student interested in writing and researching for Lawfare? Join our Student Contributor Program for the 2026-2027 academic year!
01615
Reposted by Andrea Matwyshyn
A Whole Skeptic @awholeskeptic.bsky.social · 27/08/2026
Modern life can be challenging. But let us not forget; the internet has brought many wonders into our living rooms.
06517
Andrea Matwyshyn @andreamm.bsky.social · 22/08/2026
Optimist: The cup is half full Pessimist: The cup is half empty Legal Scholar : The issue is, what is a cup?
020
Andrea Matwyshyn @andreamm.bsky.social · 21/08/2026
Less tech is often more. Cheap corded headphones don't require charging, don't expose you to Bluetooth security risks in the same ways, and don't have embedded cameras. Yet, they still do their one job of being headphones. www.engadget.com/2241639/more...
engadget.com
We have more details on Apple's camera-equipped AirPods and they are pretty dang weird - Engadget
Each earbud takes its own image, which is then synchronized for use by Visual Intelligence.
020
Andrea Matwyshyn @andreamm.bsky.social · 21/08/2026
The OG autonomous brush clearing technology - goats.
000
Andrea Matwyshyn @andreamm.bsky.social · 21/08/2026
@stephruhle.bsky.social Please do not give airtime to AI-booster doctors without disclosing their potential conflicts of interest and without including a tech expert to fact check technology assertions. We don't need another Theranos. Background reading: lawreview.law.ucdavis.edu/archives/59/...
010
Reposted by Andrea Matwyshyn
Natalie Bennett @nataliegreenpeer.bsky.social · 21/08/2026
Reminder of how little we understand even about our own bodies. We need to #Unlearn 20th-century hubris www.ft.com/content/403a...
ft.com
Blood clot discovery upends consensus about how wounds heal
Proteins stack up in layers like sheets of paper to start to form clots, scientists find
24912
Andrea Matwyshyn @andreamm.bsky.social · 20/08/2026
Tracking AI product capabilities versus human children's capabilities is a useful benchmark imho. Counting and math have been recurring failure points in the current generation of AI products. So have basic real-world interactions/ playing well with others. www.tiktok.com/@huskistaken...
tiktok.com
Maybe a 4th is needed
TikTok video by Husk
100
Andrea Matwyshyn @andreamm.bsky.social · 20/08/2026
@velshi.com Appreciated last night's discussion of data centers. But, the "AI race with China" framing is more lobbyist spin than accurate scoping. The real natsec crisis: computer security basics are broken. Fixes don't map to current data center buildouts. www.lawfaremedia.org/article/it-s...
lawfaremedia.org
It’s Morning Again in Pennsylvania: Rebooting Computer Security Through a Bureau of Technology Safety
In order to escape the computer security bootloop, Congress can create a new technology safety regulator of last resort—the Bureau of Technology Safety (BoTS).
000
Andrea Matwyshyn @andreamm.bsky.social · 30/07/2026
Welcome to the future in all its (involuntarily loud) "optimized" glory. 😑
A sign that explains that the TV volume is controlled by the remote corporate office and cannot be altered. The giant screen with the previously-referenced sign.
130
Andrea Matwyshyn @andreamm.bsky.social · 12/07/2026
The audiobook for Prospecting for Progress is now up: www.andreamm.com/progress
andreamm.com
Progress | Andrea M. Matwyshyn
010
Andrea Matwyshyn @andreamm.bsky.social · 10/07/2026
The noble capybara, a friend to all and, imho, inspo for the future of tech regulatory design. See slide 43: www.linkedin.com/posts/amatwy...
020
Andrea Matwyshyn @andreamm.bsky.social · 10/07/2026
Are you looking for a funded PhD opportunity? Are you interested in neural tech, law, and human rights? If so, here's an *amazing* one. www.linkedin.com/posts/amatwy...
linkedin.com
Are you looking for a funded PhD opportunity? Are you interested in neural tech, law, and human rights? If so, here's an *amazing* one. PhD in Law Opportunity with Professor Jennifer A Chandler…...
Are you looking for a funded PhD opportunity? Are you interested in neural tech, law, and human rights? If so, here's an *amazing* one. PhD in Law Opportunity with Professor Jennifer A Chandler, Uni...
020
Andrea Matwyshyn @andreamm.bsky.social · 01/06/2026
Too few opossums.
010
Andrea Matwyshyn @andreamm.bsky.social · 01/06/2026
Amazing! Thank you! You will be thanked in Footnote A of my article, and I will share it with you when it's out, in case it's of interest. One of your other tweets is also cited in the piece. Thanks again!
010
Andrea Matwyshyn @andreamm.bsky.social · 01/06/2026
Ooh! That is exactly what I was looking for! Thank you! Are the archives online or IRL? Would you have a pointer to the full citation (this is for an article)? If IRL, I can drop by this week. Is there a definitive Ohio sewer history book? I can't believe there was low public support in 1881!
110
Andrea Matwyshyn @andreamm.bsky.social · 31/05/2026
@neorsd.org I'm a fan of sewer and Cuyahoga River (on fire) history. Has Ohio ever had a sewer issue equivalent to the London Great Stink of 1858? If not, do you have a favorite comparably stinky (and toxic) US sewer drama? Thanks!
130
Andrea Matwyshyn @andreamm.bsky.social · 30/05/2026
@velshi.com Appreciated your piece on AI and data centers today. But AI isn't "smarter" than humans, just faster; the AI biz hype is based on false assumptions. You might find my recent law review article Exploit Machina of interest, in particular Part III. lawreview.law.ucdavis.edu/archives/59/...
lawreview.law.ucdavis.edu
Exploit Machina | UC Davis Law Review
110
Andrea Matwyshyn @andreamm.bsky.social · 30/05/2026
What do I think the answer is for preventing tech history from unproductively repeating itself? See pages 1758 et seq. for the federal part of my answer. TL;DR: A new Bureau of Technology Safety (BoTS) with a futures/ policy branch + a pilot projects branch lawreview.law.ucdavis.edu/sites/g/file...
Screenshot of first page of my Exploit Machina article.
000
Andrea Matwyshyn @andreamm.bsky.social · 30/05/2026
Those who think that AI can replace people who know industry history are doomed to repeat it.
010
Andrea Matwyshyn @andreamm.bsky.social · 30/05/2026
Only in the metaphorical legal sense. Though I think you're already managing to do that quite nicely in your latest adventures. :)
010
Reposted by Andrea Matwyshyn
michael veale @michae.lv · 30/05/2026
Meta's latest attempt to stop the progress of my 2018 GDPR access complaint dramatically failed in the Irish High Court last week www.irishtimes.com/business/202...
irishtimes.com
Meta loses High Court challenge to possible €430m fine
Data Protection Commission has power in inquiry into individual complaint to make orders affecting general data access practices, High Court rules
527776
Andrea Matwyshyn @andreamm.bsky.social · 30/05/2026
Unfortunately, it looks like it's time for me to repost my article about the relationship between First Amendment protection and security research (again). IYKYK. scholarlycommons.law.northwestern.edu/nulr/vol107/...
scholarlycommons.law.northwestern.edu
Hacking Speech: Informational Speech and the First Amendment
By Andrea M. Matwyshyn, Published on 01/19/15
165
Reposted by Andrea Matwyshyn
NE Ohio Regional Sewer District @neorsd.org · 17/04/2026
sewers are data centers.
i chat*GPT**
all the time.

*about
**the Gallons of Pollution we Treat

Northeast Ohio Regional Sewer District
845051
Andrea Matwyshyn @andreamm.bsky.social · 04/04/2026
What a nice photo! I'm sad to hear that he's no longer with us. They're such small creatures, but they take up such a big place in our hearts 😢
020
Reposted by Andrea Matwyshyn
WeRateDogs @weratedogs.com · 25/03/2026
A viral dog rescue from 2016 is now immortalized with a statue in Kazakhstan. When a dog fell into the Sayran reservoir, bystanders formed a human chain to pull him to safety. The statue is a reminder of the value of unity, solidarity and collective action. 14/10 for all
618354528124
Reposted by Andrea Matwyshyn
Angela Watercutter @waterslicer.bsky.social · 24/03/2026
Another exciting—and slightly terrifying—book excerpt on @wired.com today: A chapter from @profferguson.bsky.social’s new book ‘Your Data Will Be Used Against You: Policing in the Age of Self-Surveillance’. Read it here. www.wired.com/story/book-e...
wired.com
Your Body Is Betraying Your Right to Privacy
Attachment to smart devices and biometric surveillance leaves Americans more vulnerable to police searches than ever. Left unchecked it will only get worse.
14321
Andrea Matwyshyn @andreamm.bsky.social · 25/03/2026
I expected this CA verdict. Courts/juries are likely to look to a three part test - CHI: the Context, its sensitivity, and the Control of defendant over it; the nature of the Harm and the extent of its irreparability; and Intent/knowledge. More here, p. 1759 lawreview.law.ucdavis.edu/sites/g/file...
lawreview.law.ucdavis.edu
011
Andrea Matwyshyn @andreamm.bsky.social · 25/03/2026
"Meta executives knew their products harmed children, disregarded warnings from their own employees, and lied to the public about what they knew,” New Mexico Attorney General Raúl Torrez said in a press release."
010
Andrea Matwyshyn @andreamm.bsky.social · 25/03/2026
That CA verdict comes on the heels of yesterday's NM verdict: "The verdict stipulates that Meta must pay $375 million in civil penalties for violating New Mexico’s Unfair Practices Act, a consumer protection law." www.politico.com/news/2026/03...
politico.com
Meta hit with $375M verdict in New Mexico child safety case
The verdict marks a win for state regulators targeting tech platforms over harms to young users, as Meta vows to appeal.
100
Andrea Matwyshyn @andreamm.bsky.social · 25/03/2026
"The jury stated that Meta's and YouTube's negligence were a substantial factor in causing harm to the plaintiff...and that the companies failed to adequately warn users of the dangers of Instagram (Meta's platform) and YouTube (which is owned by Google)." www.nbcnews.com/tech/tech-ne...
nbcnews.com
Jury finds Meta and YouTube negligent in landmark lawsuit on social media safety
The jurors awarded the plaintiff $3 million in damages, finding Meta 70% responsible for harm caused to her, and YouTube responsible for 30%.
201
Andrea Matwyshyn @andreamm.bsky.social · 25/03/2026
Interested in tech procurement? Here's a new piece from me: www.andreamm.com/_files/ugd/2... It's part of a symposium coming up on April 17 at UVA Law. You can read all the papers here: www.vjolt.org
andreamm.com
122
Andrea Matwyshyn @andreamm.bsky.social · 21/03/2026
"Using ChatGPT and failing to check the case citations it provides constitute not only a lapse in professional judgment, but also a violation of Federal Rule of Civil Procedure 11."
000
Andrea Matwyshyn @andreamm.bsky.social · 07/03/2026
If you have specific questions, happy to give you my two cents fwiw. Mine have both been Flemish giants house rabbits and free roam...and sometimes outsmarted me. But I respect that.
110