Friday, February 27, 2026

IRS broke law ‘approximately 42,695 times’ in giving DHS data

 

Judge: IRS broke law ‘approximately 42,695 times’ in giving DHS data

Washington Post: “A federal judge has found that the Internal Revenue Service violated federal law “approximately 42,695 times” when it shared confidential taxpayer addresses. The judge wrote that the vast majority of the nearly 47,300 taxpayer addresses the IRS shared with Immigration and Customs Enforcement in August were disclosed without the IRS confirming that ICE provided a valid address for the person whose records it was seeking…


A federal judge has found that the Internal Revenue Service violated federal law “approximately 42,695 times” when it shared confidential taxpayer addresses with immigration enforcement officials last summer.


U.S. District Judge Colleen Kollar-Kotelly issued the ruling Thursday as part of ongoing litigation over a data-sharing arrangement between the IRS and the Department of Homeland Security.


Federal law requires that before the IRS hands over a taxpayer’s address, a requesting agency must first provide the IRS with the name and address of the person it’s looking for. The requirement exists to ensure that the government can access confidential tax records only for individuals it has already specifically identified.


The ruling finds that DHS did not follow this law. The judge wrote that the vast majority of the nearly 47,300 taxpayer addresses the IRS shared with Immigration and Customs Enforcement in August were disclosed without the IRS confirming that ICE provided a valid address for the person whose records it was seeking.


The IRS violated the [Internal Revenue Code] approximately 42,695 times by disclosing last known taxpayer addresses to ICE ... without confirming that ICE’s request set forth the ‘address of the taxpayer with respect to whom the requested return information relate[d],’” the judge’s opinion stated.


The case is now before the U.S. Court of Appeals for the D.C. Circuit, where the government is appealing Kollar-Kotelly’s November order that blocked the data-sharing arrangement.


DHS officials have defended the data-sharing agreement as necessary to crack down on illegal immigration.


Information sharing across agencies is essential to identify who is in our country, including violent criminals, determine what public safety and terror threats may exist so we can neutralize them, scrub these individuals from voter rolls, and identify what public benefits these aliens are using at taxpayer expense,” DHS said in a prior statement.


Taxpayer advocates have raised concerns about the data-sharing agreement since it began. Romo’s declaration stated that in thousands of cases, ICE’s requests contained addresses that were incomplete — featuring entries such as “Failed to Provide,” “Unknown Address,” or simply “NA NA.” In other cases, addresses were missing street names or street numbers. In still others, ICE listed jails, detention facilities or prisons as a taxpayer’s address without including the building’s street location.


In Thursday’s opinion, the judge used pointed language to describe the IRS’s verification standard. Under the government’s process, she wrote, ICE could have submitted a request with an address like “Don’t Care 12345” or simply “00000” and still received a taxpayer’s home address from the agencyA Treasury Department spokesman did not return a request for comment.


Thisconfirms what we’ve been saying all along: that the IRS has an unlawful policy that violates the Internal Revenue Code’s protections by releasing these addresses in a way that violates the law’s requirements,” said Nina Olson, founder of the Center for Taxpayer Rights, which has sued the government over the disclosures.


Olson emphasized that there was no precedent to her knowledge of a judge finding tens of thousands of simultaneous violations of federal law regarding taxpayer confidentiality.


I don’t know of any opinion about the IRS like this,” she added. “The kinds of mass requests that are coming in are unprecedented.”


Meryl Kornfield contributed to this report.



EU Parliament blocks AI tools over cyber, privacy fears

Breakthrough Discovery Targets Virus Infecting 95% of the World’s Population SciTech Daily



These video doorbells don’t rely on the cloud or subscriptions

Follow up to No One, Including Our Furry Friends, Will Be Safer in Ring’s Surveillance Nightmare – See Also How to Geek – “Picking a video doorbell that doesn’t rely on the cloud means you can save footage locally and not rely on pricey subscriptions. Pair it with a smart home platform like Home Assistant, and your doorbell will keep working when the internet doesn’t. Here are some ideas for video doorbells that work offline, even if they do have optional cloud subscriptions…”


  Inside the Debacle That Led to the Closure of El Paso’s Airspace

NYT
 OpenAI is Making the Mistakes Facebook Made. I quit.
Zoe Hitzig
 America Isn't Ready for What AI Will Do to Jobs
The Boston Globe
 ChatGPT's Memory Feature Supercharges Prompt Injection
DarkReading
 Lawsuit against Tesla reveals harrowing 911 call as driver trapped in burning car
The Boston Globe
 Hackers Publish Personal Information Stolen During Harvard, UPenn Data Breaches
Lorenzo Franceschi-Bicchierai
 European Commission Breached
Tom Allen
 When Prison Body Scanners Mistake Tampons and Piercings for Contraband
NYTimes
 Look for a citation
WSJ
 Risks of naive AI
Rob Slade with picky PGN comments
 Re: New Site Lets AI Rent Human Bodies
Martin Ward
 Dave Farber passed away at 91
sundry
 Info on RISKS (comp.risks)
 
 What happens to a car when the company behind its software goes under?
ArsTechnica
 Bad News—CVE 21858
from Bruce's CRYPTOGRAM
 University of Mississippi Medical Center Suffers Cyberattack, Closes All Clinics, Cancels Services
Mississippi Free Press
 A Wave of Unexplained Bot Traffic Is Sweeping the Web
Wired
 Defense Dept. and Anthropic Square Off in Dispute Over AI Safety
The NYTimes
 AI 'Arms Race' Risks Human Extinction, Warns Top Computing Expert
Barron's
 I hacked ChatGPT and Google's AI—and it only took 20 minutes
BBC)
 EU Parliament blocks AI tools over cyber, privacy fears
Politico
 Why an AI Video of Tom Cruise Battling Brad Pitt Spooked Hollywood
NYTimes
 How dark web agent spotted bedroom wall clue to rescue girl from years of harm
BBC
 Mark Zuckerberg to testify in landmark trial alleging that social media harms children
CBC
 What TikTok's Pixel Knows About Your Cancer, Fertility, and Mental Health Crisis
Disconnect
 Redefining Zero Knowledge
ArsTechnica
 AI must foster 'maternal instincts' or we risk extinction, warns Geoffrey Hinton
CBC
 Southern California air board rejected pollution rules after AI-generated flood of comments
LA Times
 AI discussion
Bill Maher
 Bezos vs. Musk: The New Billionaire Battle for the Moon
WSJ
 DoT's vibe-regulate U.S. transport with Gemini
Pivot to AI
 Seven Billion Reasons for Facebook to Abandon its Face Recognition Plans
Electronic Frontier Foundation
 CISA 2025 Year in Review
via Monty Solomon
 DHS found to have massive lying about immigrants on its web site, claims it was a " glitch"
CNN
 Dr Hilary Cass of the Cass Report has been referred to the GMC
Dr Webberly Responds
 Re: Look for a citation
Bob Rahe
 Info on RISKS (comp.risks)

From AI tools to Prince Andrew’s arrest: How newsrooms are digging into the Jeffrey Epstein files

Sydney is flooded as El Nina and Bad Bunny invade the harbour


Gold Coast couple accused of breaching freezing orders a second time amid alleged NDIS fraud investigation


Court Having Trouble Assembling Jury for Elon Musk Because People Hate Him So Much Futurism


Scientology, Jehovah's Witnesses and Revival Centre among alleged high-control charity groups


Do You Trust Epstein Elite With Nuclear Weapons?

No arms limitation treaties, few diplomatic channels, AI integration into nuclear command systems, and an elite with Epstein social values. What could go wrong?



From AI tools to Prince Andrew’s arrest: How newsrooms are digging into the Jeffrey Epstein files

Reuters Institute: “Over 3.5 million documents, 180,000 images and 2,000 videos. The Jeffrey Epstein files, released by the US Department of Justice (DOJ) in several tranches, constituted a disclosure of rare magnitude. 

This trove of documents opened a window into the ecosystem surrounding a powerful, well-connected convicted child sex offender.  The release offered journalists an opportunity to interrogate a sprawling evidentiary record and trace networks of access and influence stretching across politics, academia, finance and royalty. 

So far, journalists have broken stories on Epstein’s connections to powerful figures such as Peter MandelsonNoam ChomskySteve BannonSultan Ahmed bin Sulayem, and many others. Yet those revelations account for only a small fraction of what the files contain. As reporting teams continue to excavate the archive, more disclosures are almost certainly to come. But how are journalists identifying patterns of power and proximity in such a huge trove? What, precisely, are they looking for? 

And how do they search for it?  To answer those questions, I spoke with five editors and newsroom leaders from the BBC, the New York Times, the Guardian, the Miami Herald and Bellingcat, who are coordinating coverage of the Epstein files in their newsrooms across multiple beats…”

See also AI-Powered Epstein Files Research Databases and Data Explorers


How to buy a law firm if you’re not allowed to buy a law firm

FT.com – no paywall – “…So, how exactly can you buy a law firm if you are not allowed to buy a law firm? The answer is that the law firm gets split into two parts: one entity, fully controlled by lawyers, that provides legal advice to clients, and a second, called a management services organisation, that houses technology and other back-office assets and all the non-lawyer staff. The MSO sells services back to the other side of the business in exchange for fees that are large enough to make the entity profitable, and it is the MSO that sells an ownership stake to private equity. 

That way, private equity does not fall foul of the American Bar Association’s Rule 5.4, which says non-lawyers cannot be in the business of law and firms cannot share fees with non-lawyers. “Investors are trying to fine-tune the structure and get a handle on what exactly constitutes legal work,” said Austin Maloney, a lawyer at Hunton Andrews Kurth, which advises private equity. “The goal is to perform as many services as possible in the MSO. 

No investor is going to sign up to this if they can’t get the right amount of juice out of it.” The MSO structure has been used by a handful of small US firms over the past 20 years, and concentrated in those specialising in personal injury cases, but interest has exploded in recent months. There appear to be several reasons. Private equity has swept through the US accounting sector in the past five years, leaving law the last frontier in professional services. 

Meanwhile, a handful of states, notably Arizona, have set aside Rule 5.4 to allow new ownership structures breeding an openness to experimental financial models. A Texas state bar committee ruling last year effectively gave their blessing to the MSO structure, as long as it leaves control over legal cases in lawyers’ hands and the financial payments to the MSO are not calculated as a share of legal fees.

Private equity and their advisers are beginning to get confident about extracting significant amounts of a law firm’s value by shifting more assets into the MSO, potentially expanding the number of firms that could make a deal add up. “A lot of the discussion and creativity comes up in the fee structure by which the firm pays the MSO,” said Lucian Pera, a professional ethics lawyer who says he advised on a law firm MSO as long ago as 2006 and has never been busier than now. In MSOs set up by personal injury law firms, for example, Pera said the rights to the name and likeness of the lead attorneys can be owned by the MSO and licensed back. Dealmakers might hire valuation experts to weigh in on what is a fair fee. As long as the rates are fixed or calculated on a per-lawyer basis, rather than linked to firm revenue, it will not fall foul of Rule 5.4’s ban on fee-sharing, he said. “We value everything step by step and charge it back to the law firm at a market rate,” said Seth Deutsch, founder of Samson Partners, an advisory firm that has published a “how to” guide for law firm founders wanting to sell their business and has helped structure MSO deals…”

 

Microsoft Says Bug Causes Copilot To Summarize Confidential Emails

“Microsoft says a Microsoft 365 Copilot bug has been causing the AI assistant to summarize confidential emails since late January, bypassing data loss prevention (DLP) policies that organizations rely on to protect sensitive information. 


According to a service alert seen by BleepingComputer, this bug (tracked under CW1226324 and first detected on January 21) affects the Copilot “work tab” chat feature, which incorrectly reads and summarizes emails stored in users’ Sent Items and Drafts folders, including messages that carry confidentiality labels explicitly designed to restrict access by automated tools. 


Copilot Chat (short for Microsoft 365 Copilot Chat) is the company’s AI-powered, content-aware chat that lets users interact with AI agents. Microsoft began rolling out Copilot Chat to Word, Excel, PowerPoint, Outlook, and OneNote for paying Microsoft 365 business customers in September 2025…”


Epsteinalysis.com

Under the moniker – Axiomofinfinity – for which there is no further information that I could locate – a individual or group has posted a remarkable searchable database, Epstein Files Explorer, of over one million documents and over two million pages that comprise the Epstein Files released by the DOJ. Applications used – Programmatic applications – Extracted via spaCy NER and clustered by similarity. = curated key figure. Use “Known Only” to filter.

Features include the following:

  • Search – There is a search feature for: Documents, Bates numbers, Entities
  • Analyze – Timeline; Event, Network, Events, Meetings
  • Images – All Images, Faces
  • Videos – All Videos
  • Redactions – Inconsistencies, Statistics


Trump Action Tracker

Trump Action Tracker – “This project is led by Professor Christina Pagel, a UK-based health services researcher and science communicator. 

She created and maintains the dataset underlying the tracker, drawing on her background in data analysis and evidence-based policy to monitor and document how the trajectory of Trump’s presidency is aligned with the authoritarian playbook. There are a few different types of actions included in this domain:

  • Any action weakening democratic checks and balances such as restricting press freedom, undermining the authority or oversight of Congress, reducing the independence of arms length bodies, violating political norms, making it harder to hold power to account, or restricting access to public information
  • Any action undermining States’ rights
  • Any action that undermines the Rule of Law, such as violating court orders or the Constitution, circumventing Congressional appropriated funds, attacking law firms or judges.
  • Hollowing state / weakening federal institutions
  • Suppressing dissent / Weaponising state power against ‘enemies’
  • Controlling information including spreading misinformation and propaganda
  • Control of science to align with state ideology
  • Attacking universities, schools, museums, culture
  • Dismantling Social Protections & Rights
  • Corruption and Enrichment
  • Aggressive Foreign Policy & Global Destabilisation
  • Anti-immigrant or Militarised Nationalism

Thursday, February 26, 2026

The political effects of X’s feed algorithm

 “The man who moves a mountain begins by carrying away small stones.” – Confucius


“It does not matter how slowly you go as long as you do not stop.” – Confucius



Down Under Bad Bunny Touches Down in Sydney – and Heads Straight for a breakfast at Darlinghurst Cafe - The Rusty Rabbit … 


The political effects of X’s feed algorithm

Gauthier, G., Hodler, R., Widmer, P. et al. The political effects of X’s feed algorithmNature(2026). https://doi.org/10.1038/s41586-026-10098-2 – The political effects of X’s feed algorithm. Feed algorithms are widely suspected to influence political attitudes. 


However, previous evidence from switching off the algorithm on Meta platforms found no political effects. Here we present results from a 2023 field experiment on Elon Musk’s platform X shedding light on this puzzle. We assigned active US-based users randomly to either an algorithmic or a chronological feed for 7 weeks, measuring political attitudes and online behaviour. 


Switching from a chronological to an algorithmic feed increased engagement and shifted political opinion towards more conservative positions, particularly regarding policy priorities, perceptions of criminal investigations into Donald Trump and views on the war in Ukraine. In contrast, switching from the algorithmic to the chronological feed had no comparable effects. Neither switching the algorithm on nor switching it off significantly affected affective polarization or self-reported partisanship. To investigate the mechanism, we analysed users’ feed content and behaviour. We found that the algorithm promotes conservative content and demotes posts by traditional media. 


Exposure to algorithmic content leads users to follow conservative political activist accounts, which they continue to follow even after switching off the algorithm, helping explain the asymmetry in effects. These results suggest that initial exposure to X’s algorithm has persistent effects on users’ current political attitudes and account-following behaviour, even in the absence of a detectable effect on partisanship.


DOJ Scrubs Record of Interviews With Trump Accuser From Epstein Files

Follow up to Epsteinalysis.com via The New Republic: “The Department of Justice spoke four separate times to a woman who credibly accused Donald Trump of having sex with a minor he met through Jeffrey Epstein—but most accusations against the president appear to have been removed from the government’s documents on the alleged sex trafficker. 


21-page slideshow buried in the massive trove of Epstein-related documents included allegations that sometime between 1983 and 1985, 


Trump forced a woman to give him oral sex when she was in her early teens. When the woman bit down on Trump’s exposed penis, he allegedly punched her in the head and kicked her out. That same woman told the DOJ that Epstein had introduced her to Trump in 1984…”

Public Image and Sex Pistol

Happiness? That’s nothing more than good health and a poor memory. 
– Albert Schweitzer






Public Image" is the debut single by post-punk band Public Image Ltd (PiL), released in October 1978 by former Sex Pistols frontman John Lydon. It serves as a direct, defiant response to his time in the Sex Pistols, addressing feelings of exploitation by management and the press.
  • Release & Charting: Released on 13 October 1978, the single peaked at number 9 on the UK Singles Chart on 4 November 1978.
  • Context: Written by Lydon (formerly Johnny Rotten) following the 1978 breakup of the Sex Pistols, the song serves as a statement of self-determination.
  • Sound: It features a, heavy, bass-driven post-punk sound, distinctly different from the punk style of the Sex Pistols.
  • Themes: The lyrics address the commodification of Lydon's image, specifically criticizing Sex Pistols manager Malcolm McLaren.
  • Members: The original lineup included John Lydon (vocals), Keith Levene (guitar), Jah Wobble (bass), and Jim Walker (drums).
 

What can we do about corruption?

I hit the corruption theme in The National this morning: As is my habit, the angle of approach in that article differs from what I
Read the full article…


"It's pretty wild": Marc Fennell on the mystery of fraudster John Friedrich


In 'Australia's Greatest Conman?' Marc Fennell sets out to fill in the blanks in the story of the high-profile mystery man who fooled almost everyone. He talks to SBS about why this was a fraud unlike most others.


But Fennell wanted a more personal way into this one. “For me, Australia’s Greatest Conman was always about the person.”

That person was the so-called John Friedrich, a mysterious man who seemingly popped up out of nowhere in the ’80s, with no backstory to speak of, and got involved on the ground floor of the Victorian Division of the National Safety Council of Australia (NSCA).