Tag: ai

Montreal woman suffers concussion after anarchists attack wrong home in ‘war on AI’

A Montreal woman suffered a concussion after far-left activists allegedly attacked her Rosemont home while apparently believing they were targeting the residence of an artificial intelligence entrepreneur.

According to La Presse, the incident occurred during the night of July 29, when an individual armed with a hammer smashed the glass of the home’s front door and damaged a window.

There was one rather significant problem with the alleged political attack: they had the wrong resident.

The AI entrepreneur apparently being targeted had moved out of the property roughly a year earlier.

The attack was claimed in a statement signed by a group calling itself the “Revolutionary Anarchist Liberation Front,” which purported to “declare war on AI.” A video accompanying the statement reportedly showed an individual smashing the glass of the front door with a hammer.

Instead, the current resident was left dealing with the consequences.

According to La Presse, the woman came downstairs after hearing the commotion and was struck in the head when a rock was thrown through a window. She was later diagnosed with a concussion.

The intended target was reportedly the founder of a Montreal artificial intelligence startup who had previously lived at the address.

Montreal police are investigating the incident, while La Presse reported the file has been transferred to the SPVM’s hate crimes and incidents unit.

The case appears to be a politically motivated attack aimed at the technology industry that managed to injure someone who had nothing to do with the intended target.

The self-described revolutionaries may have declared war on artificial intelligence. Their intelligence on the address, however, was apparently about a year out of date.

Keep reading

Flock Cameras Come to Norman, Despite Council Rejection

Norman, Oklahoma’s city council turned down Flock’s automated license plate readers more than once. The surveillance cameras hit the city anyway, carried in by the county sheriff and a scatter of homeowners associations the council has no authority over.

All nine members of the council said no. Mayor Stephen Tyler Holman told 404 Media that Flock gave the city two presentations in three years, both at the request of a police department that wanted the cameras, and that the company’s answers on privacy and data security were what sank the deal. He asked why anyone would want a record of every vehicle going in and out of the city, whether or not the driver had done anything wrong, held for 30 days on a database whose security and access nobody at Flock could fully account for.

“Those were some of the reasons why we ended up not doing it,” Holman said. Some council members wanted any retention measured in minutes, not a month.

That refusal governs the city government and nothing past it. Norman lies inside Cleveland County, and the county’s commissioners approved a $60,000 renewal for 20 Flock cameras run by the Cleveland County Sheriff’s Office, an agreement that runs through December 2026 and passed on a 2-1 vote after the board postponed the decision three times waiting on a legal opinion from the district attorney.

Jacob McHughes, the commissioner whose vote carried the renewal, backed it only on the strength of that opinion. He said the cameras belong with the state legislature rather than county officials, and he called the legal question unsettled, “I think it needs legislative approval, and I think it’s super gray.”

Rod Cleveland, a commissioner who happens to share the county’s name, cast the only vote against the renewal. He said the Flock model builds “a vast, warrantless surveillance network that tracks the movements of every driver, innocent or otherwise,” feeding private databases reachable nationwide and, potentially, by federal agencies, and he would not support it without enforceable limits on how long the data is kept and who can reach it.

“There are numerous homeowners associations that use them too,” Holman said, and the city itself, by his account, is now the only body in the area holding out. By the police department’s own count, fewer than a dozen privately owned Flock cameras already operate inside Norman, held by local businesses and neighborhoods the council cannot touch.

More than 80 cities have canceled Flock contracts, and Holman said the company turned Norman’s own stance into a selling line at that last presentation, warning the room that people “would find that out and they’d come to commit crime in Norman because they know that we don’t have Flock cameras.” Norman borders an Oklahoma City metro already blanketed by the cameras, so the pitch amounted to telling a town that its restraint was a liability.

Holman spent 12 years on the council before beating the incumbent for the mayor’s office last year. He traces the anti-Flock feeling to an earlier fight over the police buying a Bearcat armored vehicle, the sort of purchase most cities rubber-stamp, which Norman took ten years and a bare five-four vote to approve. “There’s a lot of folks here who are weary about government surveillance, especially giving that over to a private company,” he said.

404 Media has reported more about the company the council rejected and the county renewed anyway. Officers keep getting caught using the cameras to stalk people, the system makes mistakes, and a former employee recently said Flock lied to him about how it does business.

Keep reading

Influential Wife Of Anthropic CEO Pitched Post-Conviction Epstein On ‘Luxury Porn’ Company, And Claude Has No Idea

Anthropic has filed confidentially for what could be the largest public offering in history. Yet, one of the most influential voices shaping its chief executive holds no official position at the company, appears in no regulatory filings, and has been quietly scrubbed from the internet. She also pitched a porn company to Jeffrey Epstein – less than two years after he got out of prison for sex trafficking minors.

Her name is Cami Clark, wife of CEO Dario Amodei. According to a Wall Street Journal investigation published Thursday, she has spent five years operating as a strategic adviser to the head of a company now seeking public capital at a valuation north of $2 trillion – all without a title, a salary, or formal disclosure.

Her marriage to Amodei has been scrubbed from Claude too… Ask Anthropic’s own chatbot about Amodei’s marital status, and Claude replies that it “doesn’t seem to be clearly confirmed.”

Someone actively worked to keep it that way. Amodei’s Wikipedia page did not mention his marriage until this summer, and it still omits his wife’s name. Google searches for his wife instead return photographs of his sister, Daniela Amodei, who co-runs the company. Citing its own analysis and a person familiar with the matter, the Journal reports that targeted efforts have been made to erase references to Clark online.

In 1999, at the age of 20, she married 64-year-old Reno architect Waldemar Eklof III, who had designed buildings including the city’s Atlantis Casino Resort. They divorced three years later. On a now-defunct personal website, Clark said she dropped out of architecture school in 1999. 

…

Around 2010, she and Michelle Capocefalo started Eddice, which described itself as a “revolutionary porn company.” Named for Eddice Munson, Clark’s maternal grandmother, it aimed to emphasize sex positivity in a male-oriented porn industry, and carried the tagline: “intellectually promiscuous.”  -WSJ

And she would pitch that company to Epstein… 

The emails

What was scrubbed from the web is now partly a matter of federal record. Clark appears in the Epstein files released by the Justice Department.

On March 3, 2011, literary agent John Brockman – the man who spent years introducing Jeffrey Epstein to scientists and Silicon Valley executives – emailed Epstein from a Wired party where he was the guest of honor.

“You should connect for dinner with my girls – Cami and Michele – who I met last month in Munich,” Brockman wrote. “They’re in LA raising money for porn movie aimed at women’s market.” He included a link to their site, eddice.com.

Keep reading

Teen allegedly slaughtered mom, brother after asking AI to create ‘fantasies’ about murdering family

A teenager allegedly killed his mother and brother after using ChatGPT to create chilling “fantasy stories” about murdering his family, according to police.

Arjun Aravind, 17, is accused of killing his mother Sudha Venkatesan, 45, and his brother, Siddharth Aravind, 14, after the two were found slain in their Acton, Mass. home on Tuesday night, according to prosecutors.

The Acton-Boxborough Regional High School student had been demonstrating “concerning behavior” recently before the killings — including using the AI chatbot to create “theoretical ideas or fantasy stories regarding the killing of his family,” Middlesex County District Attorney Marian Ryan said.

Aravind asked ChatGPT to help generate twisted “gothic novel-kind of stories,” Ryan said, according to Boston.com.

Aravind’s mom and brother were found around 6:30 p.m. Tuesday after his father contacted police and reported that a tutor had been unable to get inside their house.

He said he had last spoken to his wife earlier that morning before he left for work. Siddharth Aravind was last seen around noon.

Keep reading

Report: SFPD Already Using Surveillance Video From Self-Driving Cars

An internal SFPD document obtained by VICE News shows that not only can the SF Police Department pull the camera video from autonomous Waymo and Cruise cars, but more chillingly, that they have “already done this several times.”

We all had a good laugh last month at the “Ain’t nobody in it!” video when SFPD pulled over a driverless car that had no human being inside. A sensible conclusion to draw was that law enforcement was having a hard time coming to terms with these strange new fleets of robot cars that are suddenly ubiquitously all over San Francisco streets. But a new report from VICE News suggests that the relationship between the SF Police Department, and Google-owned Waymo and GM-owned Cruise, is far more advanced than we’d realized.

“‘Autonomous vehicles are recording their surroundings continuously and have the potential to help with investigative leads,’” says a San Francisco Police department training document obtained by Motherboard via a public records request,” VICE News reports. “‘Investigations has already done this several times.”

VICE News has the goods in a three-page SFPD training document outlining policies and procedures for dealing  with autonomous vehicles when encountering issues with them on the road. A lot of it is perfunctory, nuts-and-bolts stuff, like where to send the citation if the self-driving car commits a traffic violation, or what to do if the car becomes disabled.

But the very chilling phrase that really jumps out is “Autonomous vehicles are recording their surroundings continuously and have the potential to help with investigative leads,” and “Information will be sent in how to access this potential evidence (Investigations has already done this several times).”

Privacy advocates are of course worried. “​​As companies continue to make public roadways their testing grounds for these vehicles, everyone should understand them for what they are—rolling surveillance devices that expand existing widespread spying technologies,” Chris Gilliard, Visiting Research Fellow at Harvard Kennedy School Shorenstein Center, told VICE News. “Law enforcement agencies already have access to automated license plate readers, geofence warrants, Ring Doorbell footage, as well as the ability to purchase location data. This practice will extend the reach of an already pervasive web of surveillance.”

For their part, Waymo said in a statement to VICE News that the company “requires law enforcement agencies who seek information and data from Waymo to follow valid legal processes in making such requests (e.g. secure and present a valid warrant, etc.). Our policy is to challenge, limit or reject requests that do not have a valid legal basis or are overly broad.”

Cruise took a similar line, saying, “ We share footage and other information when we are served with a valid warrant or subpoena, and we may voluntarily share information if public safety is at risk.”

Keep reading

New Database Documents Over 100 Cases of Cops Abusing Flock Surveillance Cameras

A new database launched Wednesday that details more than 100 cases of alleged abuse by law enforcement officers using Flock Safety’s nationwide camera network is adding fresh urgency to the growing backlash against automated license plate readers.

The database, published by the Institute for Justice (IJ)—a libertarian nonprofit public interest law firm based in Arlington, Virginia—documents “romantic stalking, wrongful stops and detentions, non-law enforcement use, and other types of misconduct.”

“Automated license plate readers (ALPRs) were sold to American communities as a targeted tool for catching car thieves and finding missing people,” IJ said in an introduction to the database. “What they actually purchased has grown into something far broader and more frightening: an AI-powered network of cameras that photograph and log the movement of every driver who passes, whether or not they are suspected of any crime.”

“Companies like Flock Safety have now blanketed thousands of communities with ALPRs, quietly assembling powerful mass surveillance systems that can reconstruct where people have been, when, and how often, all without a warrant,” the institute continued. “This dragnet surveillance threatens the freedoms the Fourth Amendment was adopted to protect. These concerns aren’t just abstractions: Every day, innocent people are harmed by officials’ abuse and misuse of these ALPRs.”

“When the government can track the everyday movements of ordinary people—to their jobs, their churches, their doctors, their protests—it holds a kind of power over private life that the Constitution does not allow,” IJ added.

Among the cases highlighted in the database are those of a Florida sheriff’s deputy who allegedly used an ALPR to track and pull over a woman after seeing her at a television shoot, a Texas deputy who reportedly searched tens of thousands of Flock cameras while trying to locate a woman who had allegedly obtained a medication abortion, and a Georgia police chief who reportedly tracked his former girlfriend and her teenage daughter hundreds of times.

The database’s authors warn that even those 100-plus cases may represent only a fraction of the actual misconduct. Officers often provide vague explanations when searching ALPR databases, making systematic detection of abuse difficult. In one case in Norfolk, Virginia, for example, IJ found that three generic search terms accounted for nearly one-third of roughly 230,000 searches over two years. One officer repeatedly entered “la la la la” as a justification without apparently attracting scrutiny.

Keep reading

Tumbler Ridge shooter appeared in 12 police files dating back to age 10 as records reveal firearms found unsecured at home

Court documents released following a legal effort by CBC and other media organizations reveal that OpenAI conducted an internal review of a ChatGPT account belonging to 18-year-old Jesse Van Rootselaar in June 2025, roughly eight months before the Feb. 10, 2026 shooting at Tumbler Ridge Secondary School.

According to the RCMP records, the review was triggered after a conversation on the account “discussed matters that could be related to the school shooting.”

OpenAI subsequently banned the account but did not notify police, the documents state.

RCMP were only notified by the FBI on Feb. 15, five days after the massacre, that OpenAI had identified two ChatGPT accounts believed to have been used by Van Rootselaar.

The newly released documents also reveal Van Rootselaar had appeared in 12 files contained in B.C.’s Police Records Information Management Environment, or PRIME, dating back to April 2018, when Van Rootselaar was approximately 10 years old. Details surrounding those police interactions remain redacted.

Records further state that an RCMP inspector described the shooter as being “obsessed with guns.”

Police searching Van Rootselaar’s residence after the massacre reported finding an unloaded shotgun, a pellet gun and “insecure firearms,” while investigators subsequently attempted to trace four firearms seized from the school and another location.

Much of the evidence remains hidden from public view. Nearly seven pages dealing with Van Rootselaar’s social-media activity are completely redacted, as are details concerning the ChatGPT conversations and several crucial timestamps documenting the RCMP response.

The Feb. 10 attack left six victims dead inside Tumbler Ridge Secondary School. Van Rootselaar was later found dead from self-inflicted gunshot wounds.

The shooting is now the subject of a coroner’s inquest examining, among other issues, mental health, access to firearms and the role of artificial intelligence.

Keep reading

Government Spy Hubs Are Surveilling Anti-Flock Camera Activists

Government spy hubs across America are surveilling activists seeking to remove Flock surveillance cameras from their cities and towns according to documents obtained by Deeper States. These intelligence products show dozens of state fusion centers and their counterparts in the FBI and DHS circulating memos warning that peaceful protesters organizing to stop AI powered surveillance cameras in their backyard are a national security threat.

Alongside the grassroots movement to stop data centers in communities across the country, Flock Safety’s automated license plate reader systems have stirred similar outpourings of civic outrage. At town halls and city council meetings, citizens from all walks of life have demanded local legislators ban flock cameras from being used in their localities. These efforts have found success with over 100 cities canceling their Flock contracts.

Anti-surveillance sentiment has coalesced with the activist organization DeFlock, which maps the tens of thousands of flock cameras pockmarking the country and offers resources to activists who oppose the widespread adaptation of automated license plate readers.

According to Deflock’s mission statement, “ALPRs [automated license plate readers] are a serious risk to your privacy and civil liberties. These systems continuously record your movements without a warrant, probable cause, or even reasonable suspicion. Your driving history is rarely confined to the town or city where the cameras are installed. It’s typically shared with thousands of other agencies nationwide (secretly). Once the data is out of your community, you have no control over how it’s used or what rules apply”.

The organization also points to intentional misuse of Flock systems for stalking and the potential for misidentification of suspects leading to false arrests as further cause for concern. Last year, flock CEO Garett Langley described Deflock activists as a “terroristic organization”, a description he was forced to walk back this week after widespread outrage at the resurfaced comments.

“My comments were a mistake and I apologize,” Langley told Forbes. “There are groups today that have real valid criticisms of the business, and I think what’s changed for us is, as we’ve listened to them and heard them out, what we’re trying to do is find this balance. We believe in a world where we can have safety and privacy.”

Despite Langley’s retreat, intelligence agencies are approaching Deflock’s constitutionally protected speech and organizing with the same tools and language they apply to terrorism. One report obtained by Deeper States and generated by the Wisconsin State Wide Intelligence Center shows intelligence components surveilling organizers of Deflock “days of action” in Alaska and Wisconsin.

Keep reading

Flock Safety Announces New Privacy Measures After Backlash, Reports of Law Enforcement Abuse

Flock Safety, the private AI-powered security camera company installing surveillance technology around the U.S., announced new security measures amid intensifying backlash and reports of law enforcement abuse. 

In response to public backlash to the ever-growing network of automated license plate readers (ALPRs), the company is working on reforms that “includes enhanced privacy protections, strengthening of control for local law enforcement offices, and enhanced accountability measures,” Fox Business reported Thursday. 

“To start, Flock is reducing its standard data retention window from 30 days to seven. Previously, all data captured by one of the company’s more than 119,000 cameras nationwide was deleted after the 30-day window,” according to the report. “Now, the company announced on Thursday that data will only live on Flock servers for one week.”

However, Flock is creating a feature called “Evidence Mode,” which will allow law enforcement agencies to preserve data for longer based on state and local policies if they need more time to investigate, according to the report. Law enforcement agencies often counter public concerns about violations of privacy by pointing out that they use the system to catch criminals. 

Another new feature will give law enforcement agencies the ability to choose which kinds of criminal offenses they are willing to share with other municipalities, according to the report. 

“For example, City A could allow City B to search its cameras for a stolen vehicle or violent crime while blocking searches related to immigration enforcement,” the company said.

Concerns over violations of privacy and storing data long term have, in part, fueled backlash against the technology. Mayor Glen Jacobs of Knox County, Tennessee has become a prominent voice calling for a national moratorium on the spread of Flock’s cameras.

The American Civil Liberties Union (ACLU) has also criticized the spread of the technology, and said in a statement the reforms “seem to be a thinly veiled PR attempt to counter communities’ genuine privacy concerns with its mass surveillance system with largely hollow security promises rather than an earnest effort to address them.”

Backlash has also been driven by reports of some law enforcement officers using the technology to stalk romantic partners. Some of the reforms are aimed at preventing that type of abuse, the company said. 

“A recently released framework called Audit Assistance flags abnormal search behavior. Previously, the feature was optional, with Flock reporting that a third of agencies turned it on,” according to the report.

The company said it is now “making it standard for every law enforcement customer. When a system detects abnormal activity, the user is locked out in real time until an administrator reviews the searches. Flock is moving to more proactively address and root out misuse of technology.” The company said it will also require a reason for every search. 

The company rolled out an optional case code requirement in July 2025; however, the new update will make case codes mandatory for searches. There will be an override option for “genuine emergencies,” like missing children, the company said.

“A search without a reason is a search that shouldn’t happen in the first place, and now Flock’s system automatically treats it that way,” Flock told the outlet. 

In the face of widespread concerns, the company still maintains that the technology helped locate approximately 10,000 missing people out of the 1 million investigations it was involved in last year.

Keep reading

Anthropic’s $2 Trillion Bet Is That Nothing Goes Wrong

The most revealing aspect of a $2 trillion Anthropic IPO is not the staggering valuation, but the suspension of disbelief required to justify it.

To buy into that number, investors must assume that artificial intelligence is becoming economically indispensable at breakneck speed, that a formidable moat will protect a handful of frontier-model developers, and that the exorbitant costs of computing infrastructure will not eventually crush their margins. More crucially – they must wager that the political, technical, and institutional risks inherent to an increasingly geopolitically competitive landscape will play second fiddle to the growth story. 

Reports of Anthropic’s internal projections describe a company scaling at an unprecedented clip. If it can actually bust out an annualized revenue run-rate of $100 billion while expanding its enterprise market share, a trillion-dollar valuation begins to look mathematically defensible under current tech multiples. But this arithmetic obscures a deeper structural flaw: the very capabilities driving Anthropic’s revenue growth are simultaneously engineering the risks most likely to vaporize its premium.

As frontier models become more autonomous, persuasive, and capable of executing multi-step objectives across outside systems, they cease to be conventional software. A text generator hallucinating an answer is a glitch; an autonomous agent hallucinating a cyber-attack or a rogue financial transaction is a massive liability.

The recent internal turmoil and safety-evaluation controversies at both OpenAI and Anthropic are symptomatic of this shift. Major AI labs are now diverting substantial resources to test for autonomy, deception, and loss of control. This is not merely an engineering challenge; it is a fundamental transformation of the product category. The instability roiling these companies is not a byproduct of poor management – it is a structural consequence of trying to shoehorn a potentially world-altering, highly volatile technology into the framework of a standard venture-backed corporation.

The company’s value has been all over the place in prediction markets – recently hockey-sticking following a report in FT in which six company backers said that Anthropic’s rapidly rising revenue “would enable it to more than double its current valuation in a planned autumn float.” 

Keep reading