Meta has enabled Muse to build extensive and ever-growing profiles on the orbit of people in a user’s life. Based on user input but also “publicly available information.” Whether that means a web search, Facebook’s own extensive data, or buying data broker info…we don’t know? And we can’t opt out.

Muse Creates Detailed Profiles of All Your Friends and Family

Millions have downloaded Meta’s AI agent Muse. But getting it to do your bidding comes with privacy costs.

wired.com

Tyler King@tyleraking.com · 5d ago

A key component of the Cambridge Analytica outrage was your friends subjecting you to data theft without your consent or knowledge. Facebook is running the same playbook on your text messages. Even if you’re not on Meta properties, they’re ingesting your interactions.

last year we did a little made up holiday called Opt Out October that was all about taking control of your data across platforms. This year we decided to talk about how many excellent alternatives to the products made by big tech are out there www.eff.org/pages/welcom...

Welcome to Opt Out October. Let’s Take Control of Our Data and Our Devices

Big tech companies are hell-bent on taking away our agency and forcing us to use their surveillance-based services, apps, and devices. It’s hard to fight against, and it often feels much easier to giv...

eff.org

Most state attorneys general released Meta from Cambridge Analytica-related claims as part of the child safety settlement. New Mexico wasn't part of that deal, and a jury there just found 26 of 29 past statements about Cambridge Analytica, hate speech and misinformation willfully deceptive.

Eight Years On, Cambridge Analytica Scandal Catches Up With Meta in Santa Fe

A jury found that Meta—then Facebook—misled the public in statements about Cambridge Analytica and its policies on hate speech, violence and misinformation.

buff.ly

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

The Greenland deal does not substantially alter the status quo. That's good but saying too loudly that Trump backed down jeopardizes the agreement. Trump's fragile ego and his limited exposure to negative information thus present a variation on what Robert Putnam called a two-level game. Details 👇

This is insanity. Irish Presidency's draft text of amendments to the GDPR allows training & operation of any AI without safeguards & tough legal basis (under "legitimate interest" a company claims what it is doing is OK. Up to you to find out and challenge in court). This is a free pass for US Tech.

“‘Certain companies have used accessibility and disability rights as a Trojan horse for a tool that poses significant privacy and security threats,’ Center for Democracy and Technology disability lead Ariana Aboulafia told me, ‘and to be clear, the privacy concerns affect disabled people too.’”

Is Meta using disabled people to promote its glasses?

AI glasses are a double-edged sword for disabled people, who face unique risks from surveillance.

motherjones.com

The New Yorker Jesus Christ. Anyone who had even a word to say about Hunter Biden should be dropping their jaws over this.

T minus 30 days to publication of Power and Greed: Monopolies, Mergers, and Cartels on the American Stage. Learn more and preorder at www.powerandgreed.com

The hacking tools the chatbot companies are developing stand to carry on this very stupid tradition. It *is* scary, but not because the chatbots are waking up. It's scary because the world's IT systems are indifferently created and poorly maintained and riddled with vulnerabilities: xkcd.com/2347/

Dependency

xkcd.com

Excellent work. Thanks for sharing. Will be added to class reading. We spent a lot of time discussing misleading tech metaphors and distortion of legal and policy analysis

Dr Abeba Birhane@abeba.blacksky.app · 3w ago

"Did these AI models “go rogue”? Not exactly. the models were tasked with finding solutions to a hacking challenge, and this is what they attempted to do, though in an unintended and surprising manner." aiguide.substack.com/p/misleading...

Still time to register for the Cognitive Security for Human–AI Systems (CogSec) workshop in a few weeks in DC at the 14th ACM Collective Intelligence Conference and HCOMP 2026: ACM Conference on Human-AI Complementarity and Alignment programs.sigchi.org/ci/2026/prog...

Slowing Down to Allow for Emergence speculativefriction.substack.com/p/slowing-do...

Slowing Down to Allow for Emergence

Perspectives from the Frictional AI workshop at the Hybrid Human-AI Conference in Brussels

speculativefriction.substack.com

Yep. Exactly

Mar Hicks@histoftech.bsky.social · 3w ago

The way AI is going to “kill everyone” is just by manufacturing consent for ever worse outcomes & conditions, like many governments do. It is a governing technology after all. And like bureaucracy it launders responsibility so the people & institutions who deploy it aren’t accountable for the harms

No, no, no. Oh, the credulity. They don't think. They don't lie. They are neither conscious nor sentient. They calculate. Just stop. Who Cares if AI Is Conscious—It’s Basically Alive www.wired.com/story/who-ca...

MAGA has been obsessed with proving an alleged plot by Biden officials, social-media companies, and researchers to suppress conservative speech. Jobs were lost, a center was closed, and reputations were tarnished. Now we know it was all a fiction. @anneapplebaum.wsocial.eu. buff.ly/BaSs2Cr

Amazon's payola isn't about radio play - it's about *search*. When you search Amazon, the top results do not represent Amazon's best guess about what product will best match your query: rather, Amazon *auctions off* those top results to its platform sellers. 22/

AGI !!!! It’s here at last.

Sober & pointed.

Dave Karpf says I should worry more about Trump stealing votes than AI chatbots

Bret Stephens’ least favorite thinker ties the history of the future to our present collapse.

thecause.substack.com

Waymo and other AV companies decide whether or not their vehicles will follow the law. Increasingly, they are deciding not to. Regulators can and should change their calculations.

Tesla’s ‘Mad Max mode’ points to a big problem for self-driving cars

Human drivers break traffic rules all the time. To stay competitive, computer-powered cars are doing the same.

fastcompany.com

We must stop these socialists before they tell companies who they can and can't hire, threaten executives into giving the government part ownership of their companies, pick winners and losers through selective tariffs, subsidies, and regulation, and dictate private sector personnel and HR policy.

The Bulwark@thebulwark.com · last mo.

Mike Johnson: "The barbarians are now inside the gate. By our count, there's almost 40 openly avowed Marxists, Communists, Socialists. They're all in the same camp…This is a serious threat to us."

A little more technical: OpenAI failed to fully airgap their sandbox, leaving in place a repository manager that had limited access to the open internet. Agents discovered a zero day vulnerability and exploited it, which is unsurprising given the temperature setting of the models

I will explain this whole story in plain English for you: OpenAI did not fully isolate their experimental environment from the internet, accidentally leaving in place a software program that could access the open internet for updates. Agents exploited it, which is what they are structured to do.

ali alkhatib@ali-alkhatib.com · last mo.

... what the hell are these people talking about?

As with the mail-in ballot case, so in the ballroom case, SCOTUS includes a whopper in the opening line. Today: "In October 2025, the federal government started a project to replace the East Wing of the White House." "The federal government?" More like "the rogue president who thinks he's a king."

There are issues with the entire research field around AI and consciousness, but why does Cade Metz always get catfished into writing about how a 20-something founder with a bachelor's degree received the same AI spam we all do and now thinks—or finds it lucrative to pretend—computers are conscious?

Study A.I. Consciousness? The Bots Would Like a Word With You. (Gift Article)

Given access to email, A.I. agents have started reaching out to the philosophers and researchers exploring deep questions about them.

nytimes.com

Excellent piece. @mblawrence.bsky.social @ormerod.bsky.social this resonates strongly with the decomposing AI systems work

Lawfare@lawfaremedia.org · last mo.

Invisible decisions made by AI companies, like which sources are used and maintained, are having an increasing impact on our information flows. @klonick.bsky.social and @noupside.bsky.social explain how this invisible source skew happens and its impact on the information ecosystem.

The good old "the actual harms you are seeing & experiencing are a distraction from what will happen with the magical machine god (that's coming around any day now) argument." Its so interesting to see how when they see the public isn't falling for this bullshit anymore they peddle it even more.

The Atlantic@theatlantic.com · last mo.

AI systems are at risk of becoming dangerously powerful—and there may be only one way to control the technology’s development, Rogé Karma argues: theatln.tc/j3zw1Ys4 🎨: The Atlantic. Source: Getty.

Anthropic shared usage data with Stanford's SALT Lab. The Lab's study found "productive friction" when users knew & cared enough to push Claude. Good illustration for students re utility of expertise. #friction @neilturkewitz.bsky.social @brettfrischmann.bsky.social www.alphaxiv.org/abs/2608.hum...

New paper alert: Affective Context Amplifies Sycophancy in LLM Responses arxiv.org/abs/2608.21242

Affective Context Amplifies Sycophancy in LLM Responses

As conversational companions, large language models (LLMs) often have access to users' emotional states. We study how this affective context modulates LLM sycophancy in subjective, evaluative interact...

arxiv.org

Anybody who’s still advocating putting these tools anywhere near a school is not doing so despite the cognitive collapse they cause, but because of it. dx.doi.org/10.2139/ssrn...