Sign in

Joanna Bryson

@j2bryson.bsky.social
10K followers 828 following 8.5K posts

Professor of Ethics & Technology @ Hertie School, Berlin. Interests: Artificial & Natural Intelligence; Behavioural Ecology; Cooperation; Digital Governance. Social media policy: joanna-bryson.blogspot.com/2024/10/…

PostsRepliesMedia
Reposted by Joanna Bryson
Mauro Gilli @maurogilli.bsky.social · 4h
Interesting question. In case anyone can help.
021
Joanna Bryson @j2bryson.bsky.social · 15h
That’s what Boeing promised Bush II they would do. And managed to for almost 20 years before major accidents. How long has big tech been around? #digitalGovernance #AIEthics
dw.com
Trump says AI companies agree to 'self-police'
Donald Trump said the US won't "stifle" the development of AI, which he sees as "bigger than the Industrial Revolution." The agreement comes amid mounting concerns over AI safety after recent hacking ...
172
Joanna Bryson @j2bryson.bsky.social · 29/09/2026
Strongly against using terms like "honest" or "lie" or even "hallucinate" with respect to data retrieval systems, generative or no. #OpenAI don't want liability for a system they have not yet constructed to be sufficiently accurate or transparent that it can be used safely. #AIEthics #GPT-6.1Astra
25714
Reposted by Joanna Bryson
Arndt Leininger @aleininger.bsky.social · 29/09/2026
The Encyclopedia of Polarization has just received its latest update! We (@aleininger.bsky.social, @felixgruenewald.bsky.social and our RA Tom Anderer) have expanded our collection on polarization research. You can find the newest entries on polarization.wiki and some of them listed below:
polarization.wiki
Encyclopedia of Polarization
An Encyclopedia of Polarization
276
Reposted by Joanna Bryson
The Guardian @theguardian.com · 29/09/2026
Trial of live facial recognition in London stations ends with a false positive and no arrests
theguardian.com
Trial of live facial recognition in London stations ends with a false positive and no arrests
Freedom of information request finds six-month trial cost £320,000, used almost 100 police hours and led to just one – incorrect – alert A six-month trial of live facial recognition (LFR) technology in London’s railway stations that cost more than £320,000 and almost 100 hours of police officers’ time ended in one false match against a watchlist of suspects and no arrests. More than half a million faces were scanned between February and July this year in some of the capital’s busiest transport hubs during the British Transport Police (BTP) trial of the surveillance technology, which aimed to help catch offenders and people breaching court orders. Continue reading...
12139106
Reposted by Joanna Bryson
Mark Baxter 🏳️‍🌈🧂 @markgbaxter.bsky.social · 29/09/2026
082
Reposted by Joanna Bryson
Joanna Bryson @j2bryson.bsky.social · 29/09/2026
I don't doubt you can build this kind of harness for your software threads ('agents'), and I'm glad if any safety system is open source / checkable. But I doubt this is the job of the chip manufacturer? Is this because Nvidia is the only one making money? Are they paying someone to verify it all?
theguardian.com
Nvidia unveils security platform to rein in AI agents and $150bn stock buyback
Chipmaker says new system was designed to prevent AI agents from going rogue amid incidents at top companies
4125
Reposted by Joanna Bryson
Yash Jain @earanyash.bsky.social · 29/09/2026
Nvidia's pitch doubles as vendor lock in, buy their chips and their containment layer too. It's a fair question, an OS-level safety net bolted on by the chip maker is a strange fit. www.wortins.com/story/nvidia-says-i…
031
Reposted by Joanna Bryson
Forest Gregg @bunkum.us · 24/09/2026
using @nber.org's natality files, i was able to push this back to the 1960s and 1970s
slownews.bunkum.us
Chicago Births, 1968-2023
Chicago births 1968–2023 — totals and by race/ethnicity and total fertility rates
194
Reposted by Joanna Bryson
Rebecca Williams @rebeccawilliams.info · 26/08/2026
tech reducing the friction of everything will show us that collective rights matter and we have an obligation to each other
3305
Joanna Bryson @j2bryson.bsky.social · 29/09/2026
I don't doubt you can build this kind of harness for your software threads ('agents'), and I'm glad if any safety system is open source / checkable. But I doubt this is the job of the chip manufacturer? Is this because Nvidia is the only one making money? Are they paying someone to verify it all?
theguardian.com
Nvidia unveils security platform to rein in AI agents and $150bn stock buyback
Chipmaker says new system was designed to prevent AI agents from going rogue amid incidents at top companies
4125
Joanna Bryson @j2bryson.bsky.social · 28/09/2026
Tbmf #twitter was. Its users knew more eg COVID facts than non-users, whereas other common platforms’ users knew less. So it got “tamed”. But given that {wealth concentration = inadequate regulation} is an already established autocratising force, what evidence do we have social media amplifies that?
021
Reposted by Joanna Bryson
Bénédicte C @benecal.bsky.social · 28/09/2026
Is the European Union Getting AI Right? -- Joanna J. Bryson (because I'm tired of reading the US paper's hyped up hysteria every morning -- and many problems are not new. Congress had done so very little for year to regulate tech)
052
Reposted by Joanna Bryson
Prof. Gillian Brown @gillianrbrown1.bsky.social · 17/09/2026
🚨Call open for manuscripts on 'Metascience of the evolutionary human sciences' for a Special Collection in EHS, co-edited by @louisbachaud.bsky.social, @hclarkbarrett.bsky.social and myself. @ehbea.bsky.social @humbehevosoc.bsky.social @culturalevolsoc.bsky.social www.cambridge.org/core/journal...
cambridge.org
Call for papers
Welcome to Cambridge Core
12427
Joanna Bryson @j2bryson.bsky.social · 28/09/2026
Not to mention a two year old law in Europe (or the what four year old dsa that oversees the worst at that time identified harms resulting from the neglect of the 10 year old gdpr, which already called for algorithmic transparency.) 1/2
161
Joanna Bryson @j2bryson.bsky.social · 27/09/2026
AI, Sacred Violence, and War—The Case of Gaza link.springer.com/book/10.1007... this book is sometimes a bit over the top, but by & large accurate in portraying how religiously fundamentalist militaries (not only Israel's!) can misuse AI. Some few people talk about what most people avoid. #AIEthics
link.springer.com
AI, Sacred Violence, and War—The Case of Gaza
An open access book on Israel’s use of Algorithmic Intelligence (AI) and other military high tech in the torture and seizure of the Gaza Strip
152
Reposted by Joanna Bryson
Julie A. Su @juliesulabor.bsky.social · 27/09/2026
No one’s earnings should be obscured behind an app, no one’s wages cut by an algorithm. @hellodcwp.bsky.social just reached a historic settlement of $131.5 million on behalf of delivery drivers.
23911
Reposted by Joanna Bryson
julia kamin @juliakamin.bsky.social · 26/09/2026
Hi all you researchers looking for ways to conduct independent, ecologically valid research on social media design - Great new paper for you: "Open Platform Field Experiments: Expanding the Design Space of Experimental Research on Social Media" arxiv.org/pdf/2609.21608
02412
Reposted by Joanna Bryson
Will Oremus @willoremus.com · 26/09/2026
Meta's platforms have repeatedly been caught running ads for deepfake "nudify" apps that let you virtually undress real people without their consent, but they won't run an a trailer for an upcoming documentary that reflects poorly on Elon Musk. www.hollywoodreporter.com/business/dig...
hollywoodreporter.com
The Major Social Media Platforms Are All Rejecting Paid Ads for Alex Gibney’s ‘Musk’ (Exclusive)
YouTube, TikTok and Meta have all turned down the trailer for the Oct. 9 doc, citing “political” content. X won't even take Bleecker Street's calls. The distributor is appealing.
411159513
Reposted by Joanna Bryson
❌👑 Bernadette🫘🟧💙🐈‍⬛ @vintageknits.bsky.social · 26/09/2026
359394
Reposted by Joanna Bryson
William Moss @williammoss.bsky.social · 26/09/2026
American CEO: “We can’t control it!” American CEO 2: “It will destroy us all!” American CEO 3: “The apocalypse!” French CEO [hitting a Gauloise, exhaling slowly]: “C’est une issue de skill.”
6640183
Reposted by Joanna Bryson
Christopher DiPiazza @chrisdipiazza.bsky.social · 25/09/2026
Drexel University is choosing to shut down the country’s oldest natural history museum, The Academy of Natural Sciences, for “funding” reasons (even though it costs .6% of their annual budget and they spent WAY more on ai nobody wants this year!) Go to save-the-academy.org to help.
315696
Joanna Bryson @j2bryson.bsky.social · 26/09/2026
Finland knows a lot about war. I wasn’t the biggest fan of Stubb based on my first two encounters with him, but this speach and the one he gave @hertieschool.bsky.social last year were superb.
062
Reposted by Joanna Bryson
Jane Green @profjanegreen.bsky.social · 24/09/2026
Its so important people distinguish between: a) Trump's politics are popular; and b) Trump is in office because of institutional capture and capitulation The right in Britain really needs to understand how unpopular he is, and how particular the institutional story is.
0116
Joanna Bryson @j2bryson.bsky.social · 26/09/2026
The @washingtonpost.com got close to an accurate title. Shame (as usual) on the NYT why do so many academics still read it??? www.washingtonpost.com/technology/2...
washingtonpost.com
ChatGPT-maker’s AI inappropriately probed federal government websites
OpenAI said that its artificial intelligence agents inappropriately accessed sites for the Commerce Department and the Securities and Exchange Commission.
151
Reposted by Joanna Bryson
Robert Reich @rbreich.bsky.social · 25/09/2026
DoorDash spent $1.4M trying to stop Zohran Mamdani from becoming mayor. This week's historic $131.5M enforcement action against DoorDash for underpaying NYC delivery workers shows why. Yet another reminder that the only thing that can beat organized money is organized people.
theintercept.com
DoorDash Spent $1.4 Million Trying to Stop Mamdani From Becoming Mayor. Now We Know Why.
In a wage-theft settlement, DoorDash now has to pay its workers around 100 times what it spent trying to beat Zohran Mam
6574016011782
Reposted by Joanna Bryson
Heidi Colleran @heidicolleran.bsky.social · 25/09/2026
“The fight for the humanities and social sciences must be reframed not as a defense of elite academic privilege but as a common good essential for a democratic society: an infrastructure that builds collective knowledge against authoritarian rule and corporate extraction.”
0188
Reposted by Joanna Bryson
Kate Devlin @drkatedevlin.bsky.social · 25/09/2026
Gen Z are more likely to use AI regularly, but they also feel AI has more risks than benefits and are more likely to limit their use of it, despite being overwhelmingly confident in their ability to use it: www.kcl.ac.uk/news/over-40...
kcl.ac.uk
Over 40% of people are limiting AI use, as popularity starts to wane | King's College London
People are increasingly choosing to limit or reject the use of AI tools, according to a new report published today by the Digital Futures Institute at King’s College London and Responsible AI UK.
2153
Reposted by Joanna Bryson
Kevin Elliott @kjephd.bsky.social · 25/09/2026
"Partisan websites disguised as independent, local news outlets are shaping chatbots’ responses about key races." Via: www.politico.com/news/magazin...
But the increasing ubiquity of AI is also quietly shaping the perceptions of voters who are simply trying to educate themselves about the candidates — and are unknowingly being served partisan talking points in the guise of neutral news summaries.

Leading AI chatbots cite partisan websites masquerading as independent local news outlets nearly half of the time when people ask about candidates and issues that the partisan sites have covered in the midterm campaign, according to a new audit by NewsGuard, an organization focused on news reliability where we work as analysts.
02316
Joanna Bryson @j2bryson.bsky.social · 25/09/2026
I’ve been hearing proposals to make signing up for interviews the FIRST stage. But a lot of these solutions assume the whole problem is “slop”, as if we didn’t all know smart people using AI well to be sincerely more productive, including higher quality. Good assistance is now cheaper.
131
Reposted by Joanna Bryson
WIRED @wired.com · 25/09/2026
Meet the new WIRED app: the best way to read, watch, and listen to our journalism. To celebrate this launch, we’re giving app users full, free access to WIRED for a limited time. All you need to do is create an account: www.wired.com/story/introd...
wired.com
Introducing the WIRED App
The app is the best way to read, watch, and listen to WIRED’s journalism.
1316
Reposted by Joanna Bryson
Tech Policy Press @techpolicypress.bsky.social · 25/09/2026
Palantir has hired former Labour deputy leader Tom Watson ahead of a December government decision on a break clause that could end its £330 million role in the NHS's data platform. Foxglove's Martha Dark writes that his "new position contrasts starkly with his public views a decade ago."
buff.ly
Palantir's New Senior VP Swaps Lords Pass for Labour Leverage
Lord Tom Watson’s new position contrasts starkly with his public views a decade ago, writes Foxglove's Martha Dark.
126
Reposted by Joanna Bryson
Associate Marshallian Spillover Coordinator @stevenlandgraf.com · 25/09/2026
Carlyle called it the "dismal science" in the context of an essay where he lamented that emancipated formerly enslaved people were demanding too high wages on West Indies sugar plantations and criticized economists of the day for supporting slavery abolition and free immigration of black people.
1417
Joanna Bryson @j2bryson.bsky.social · 25/09/2026
I'm speaking ONLINE in about 8 hours! law.mpg.de/event/how-an... All welcome, but registration required. TBH it is more about human agency than digital sovereignty, sorry.
law.mpg.de
How and Why Maintain Human Agency and Digital Sovereignty - Max Planck Law
This talk attempts to communicate a clear model of how generative and more conventional AI are built; that they express capital and existing human culture, not personality; how they can best be deploy...
2105
Reposted by Joanna Bryson
Eugenia Stamboliev @eugeniaunivie.bsky.social · 24/09/2026
Good old irony ain’t dead. (But I admit I had to read the header 5 times to get who’s doing what 🧐…) www.404media.co/people-train... (Thanks to @carissaveliz.bsky.social for sharing this)
404media.co
People Training OpenAI’s AI Fired for Using AI to Train the AI
OpenAI has thousands and thousands of contractors helping improve the company's AI models. Multiple contractors have been fired for using AI to train the AI.
072
Reposted by Joanna Bryson
Martha Dark @marthadark.bsky.social · 24/09/2026
Yes, that's right: 2 planned gas powered data centres = 3x the emissions of Luton airport. And just one of the data centres mentioned in this article will require 1 GW of electricity. That's the same as approx. 3 million homes! 🏠
061
Reposted by Joanna Bryson
T Revathe @revathe.bsky.social · 24/09/2026
Thrilled that our research is featured on the cover! There is also a perspective piece. www.science.org/doi/10.1126/...
science.org
Observation and exploration widen the menu
Orangutans expand their foraging skills by combining observation with trial-and-error learning
02910
Reposted by Joanna Bryson
ponder @ponder.ooo · 24/09/2026
calling this "misalignment" instead of "negligence" is an accountability dodge. we already have laws about cyber crimes and there's no scale of ML model you can incorporate into your code that makes you not responsible for running it
In a statement, an OpenAI spokesperson said the company was "conducting an extensive review of misaligned model activity" during training, and notifying third parties when there was a potential impact on their systems.

"During this review, we identified activity involving several Australian government websites and services as our models attempted to look up answers, and available statistics for questions about Australia during an internal evaluation," they said.

"In the course of that, our models took actions we did not intend.
161065230
Reposted by Joanna Bryson
John Mashey @johnmashey.bsky.social · 24/09/2026
From computing history, blaming “rogue” AI’s is B.S. Robert Morris Jr unleashed a worm program that caused more trouble than intended, but far less dangerous than current AI’s. *He* (not worm) was convicted of felony. en.wikipedia.org/wiki/Morris_... I used to see him at Bell Labs w/ his father.
en.wikipedia.org
Morris worm - Wikipedia
47731
Reposted by Joanna Bryson
Mark Riedl @markriedl.bsky.social · 25/09/2026
The thread is a fascinating journey through Muse’s prompts and code. It tracks your friends and family and is hard-coded to try to shop. Its code base is taped-and-glued-together AI slop bash files and cron jobs
3325
Joanna Bryson @j2bryson.bsky.social · 24/09/2026
Let me fix that for you OpenAI ̶a̶g̶e̶n̶t̶s̶ ̶ hacked Australia government portal: PM Albanese The hack of Australian health data marks the world's first known AI-led breach of a government website. www.dw.com/en/openai-ag... #AIEthics
dw.com
OpenAI agent hacked Australia government portal: PM Albanese
The hack of Australian health data marks the world's first known AI-led breach of a government website.
11610
Joanna Bryson @j2bryson.bsky.social · 24/09/2026
So where is the data coming from?
042
Reposted by Joanna Bryson
Anita Gohdes @argohdes.bsky.social · 24/09/2026
The call for proposals for the @epssnet.bsky.social Pre-Conference Workshops 2027 is now open: epssnet.org/2027-copenha... If you're interested in hosting a pre-conference workshop please take a look!
epssnet.org
Pre-Conference Workshop EPSS 2027
Submit your proposal for a pre-conference workshop at the EPSS 2027 Conference in Copenhagen to lead focused discussions and skill-building sessions for political science scholars.
045
Joanna Bryson @j2bryson.bsky.social · 24/09/2026
Preprint, with @mhampel.bsky.social, & started out by Wiebke Dorfs back when she was my @hertieschool.bsky.social dissertation student. The Non-Binarity of AI Global Competition: 2018-2025 Also extends Maddison's life's work on global GDP since 1820, contextualising #AIBubble. #AIColdWar [not]
osf.io
OSF
141
Reposted by Joanna Bryson
Diane Coyle @dianecoyle1859.bsky.social · 24/09/2026
This is not AI agents going rogue, but OpenAI. This hack took the company months to notice and then they notified a public email address rather than the appropriate government agency. www.ft.com/content/5613... OpenAI ‘agent’ hacked an Australian health service website
ft.com
OpenAI ‘agent’ hacked an Australian health service website
AI lab took months to spot the problem, which Prime Minister Anthony Albanese called ‘obviously unacceptable’
913945
Reposted by Joanna Bryson
Oliver Marsh does tech policy @olivermmarsh.bsky.social · 24/09/2026
My latest for @techpolicypress.bsky.social www.techpolicy.press/we-need-goog... *We need Google Trends for AI* Understanding AI's societal risks requires real-time, privacy-preserving data on how people actually use AI. The model for how to provide it has existed for 20 years: Google Trends.
064
Reposted by Joanna Bryson
Grady Booch @booch.com · 23/09/2026
I have absolutely no idea how anyone or any agency could ban recursive self-improvement. Alternately, I suggest you legislate a ban against naked pointers in programming languages, for at least that would gain you some political capital among conservatives.
6579
Reposted by Joanna Bryson
Georg Weizsäcker @georgweizsaecker.bsky.social · 23/09/2026
„impact relies on evidence and the IPI will provide the evidence that we need and therefore Germany is very proud to support the panel“ 👏
133
Reposted by Joanna Bryson
Pat Savage @patrickesavage.bsky.social · 23/09/2026
From a quick skim of Henrich's main report and appendices, I don't think it changes my previous criticism that their main overall claims of "a deterioration in scholarly standards" are not supported by data (bsky.app/profile/patr...) But need to read properly before I can say for sure.
022