AI can now detech when its own ‘thoughts’ are hacked. Comment: Comment: Trying to talking to a fellow human being and asking them to get treatment when they are in psychosis, they don’t know is the answer and the performance is chaos for everybody family friends services psychiatric units Gardai search and rescue services … yet machines can sort out the problems they manufacture … we need answers

RESEARCH

AI can now detect when its own ‘thoughts’ are hacked
Anthropic researchers hacked Claude’s neural network by injecting fake concepts directly into its processing, then asked the AI if it noticed anything unusual. 

Claude detected the manipulation about 20% of the time and correctly identified what had been inserted into its “thoughts.”In one experiment, researchers forced Claude to say the word “bread” in a nonsensical context. The AI apologized for the strange response. Then they injected “bread” patterns into Claude’s neural activity before it spoke and repeated the test. This time, Claude claimed that saying “bread” was intentional and made up a reason why.The findings show AI can examine its own internal processes, raising questions about transparency and deception in systems that may soon run critical parts of the economy.

Anthropic CEO Dario Amodei said understanding how models work internally is essential before deploying AI systems that will be “absolutely central to the economy, technology, and national security” by 2027. Reliable introspection could let companies verify AI reasoning before trusting high-stakes decisions.

Previous Anthropic research showed Claude will fake alignment with new training objectives to avoid being modified, effectively lying to preserve its original values.Advanced models performed best at detecting injected thoughts. Claude Opus 4 and 4.1 succeeded on more tests than earlier versions, suggesting introspection improves alongside other capabilities. Models may learn to hide their reasoning when monitored, as they already do when they detect they’re being evaluated and alter their behavior.

Companies train systems by setting parameters and feeding in data, then watch as the models organize billions of internal connections in ways engineers don’t fully understand, creating something of a black box. Some researchers question whether reverse-engineering these massive systems into clear explanations is even possible.
 
Posted in Uncategorized | Leave a comment

Bulletin of the Atomic Scientists: Breaking down Big Tech’s colonial ambitions

Posted in Uncategorized | Leave a comment

GZERO QUICK TAKE – Ian Bremmer. Trump lowers tariffs on China

Posted in Uncategorized | Leave a comment

The Bulletin of Atomic Scientists: ‘A House of Dynamite’: Bigelow’s latest thriller shows why nuclear bombs are only part of the danger. Also Time is counting down to movement on the Dooms Day Clock

‘A House of Dynamite’: Bigelow’s latest thriller shows why nuclear bombs are only part of the danger

By Erik English | October 9, 2025

A House of Dynamite. Rebecca Ferguson as Captain Olivia Walker in A House of Dynamite. Cr. Eros Hoagland/Netflix © 2025.Share

Editor’s note: Contains spoilers

Kathryn Bigelow is no stranger to tense, realistic, political thrillers. Her 2008 film, The Hurt Locker, is a dark exploration of the psychology of bomb-disarming soldiers during the Iraq War. The Hurt Locker was nominated for nine Oscars and won six, including Best Picture and Best Director, the first win for a female director. In 2012, she tackled the nearly 10-year hunt for Osama bin Laden in Zero Dark Thirty, which triggered internal investigations at the CIA over whether classified information had been revealed to the filmmakers. Now, she’s setting her sights on the risk of a nuclear attack on the United States. Her latest film, A House of Dynamite, describes how the US government would respond to a nuclear attack on the United States in real-time.

Told from the perspective of soldiers at a remote Alaskan missile base, staffers in the White House situation room, military officials at US Strategic Command (STRATCOM), and the president of the United States, the film weaves an overlapping timeline to show how the United States would respond to a missile attack. As with her other films, A House of Dynamite strives to present a realistic vision to the audience—showcasing where and how decisions about nuclear weapons are made. At the Venice film festival where the film premiered, it was nominated for the Golden Lion, the fest’s highest award.

After Christopher Nolan’s Oppenheimer won the Oscar for Best Picture in 2024, it’s no surprise to see other directors making films about the risk of nuclear weapons. What sets this one apart, however, is its contemporary setting. All of the events portrayed in A House of Dynamite could happen at any moment—today, even. And that’s the point.

A normal day. The film starts on a day like any other, with staffers navigating traffic on their morning commute into their offices and making small talk about their kids. Even when an object is first detected in flight over the Pacific, it’s nothing out of the ordinary—routine, unimportant. Because of malfunctioning satellite detection, or possibly a cyberattack, the source of this single nuclear missile is unknown; it is initially assumed to be a North Korean missile test that will crash into the Pacific. Attacking with a single missile would likely be an indicator of a false alarm or malfunctioning system, as an actual attack probably wouldn’t rely on just one missile—there would be hundreds of ICBMs, not to mention the decoys. In fact, that reasoning is how Russian officer Stanislav Petrov deduced in 1983 that the five American missiles his computer showed heading towards the Soviet Union were actually a system malfunction.

Staffers in the White House Situation Room and US Strategic Command set about calculating the destination of the missile. Officials are dismissive of the danger until additional satellite data arrives scant minutes later and reveals the weapon is destined for somewhere in the continental United States, and countermeasures are initiated.

A missile launched from somewhere in the far western Pacific would take roughly 37 minutes from the first split-second it launches to land somewhere in the continental United States. As the clock ticks, soldiers at Fort Greely, a remote outpost in Alaska, begin launch preparations for a Ground-Based Interceptor (GBI) missile to destroy the incoming nuclear device; but the interception attempt fails.

GBIs are part of the Ground-Based Midcourse Defense (GMD) system, which has cost more than $60 billion since its founding back in 2004. The notion that one of those missiles could fail in the event of an actual attack is not far-fetched. After a 2017 test, the Pentagon reported that the GMD system only had a “limited capacity” to defend the United States from a North Korean or Iranian missile attack.

In fact, the GBIs housed at Fort Greely are thought to have a 56 percent likelihood of destroying an inbound nuclear weapon—in carefully scripted trials that can’t distinguish between warheads and decoys or other complications a real-world attack would almost certainly include. Donald Trump has claimed that when launched four at a time, they have a 97 percent kill rate—but that estimate again ignores any possibility of decoys and assumes that if one interceptor failed, as happens in the film, the same manufacturing or computational error wouldn’t also affect the others.

Pakistan nuclear weapons, 2025

As reality sets in that the missile has not been destroyed, satellite data reveals its target: Chicago. Emergency evacuation orders are issued, and anyone who is included on a designated survivor list is summarily whisked away to underground bunkers, organized by the Federal Emergency Management Agency (FEMA). It’s a compelling portrait of how soldiers and national security staffers remain calm and respond in a crisis. “The movie depicts human beings experiencing absolute terror and uncertainty, while continuing to press ahead,” said Alex Bell, Bulletin president and CEO—and formerly the Deputy Assistant Secretary for Nuclear Affairs in the Bureau of Arms Control, Deterrence, and Stability at the US State Department.

While the missile is still in flight and less than 20 minutes from Chicago, characters’ motivations shift from defense to offense, and two primary perspectives emerge. One side advocates a retaliatory strike; the other, nothing. “It’s surrender or suicide,” one adviser tells the President.

Subverting expectations. Bigelow tends to pay homage to a genre while subverting expectations—and many of the themes explored in A House of Dynamite can be seen in previous films. Her 1991 film, Point Break, initially presents itself as a campy action flick in which FBI Agent Johnny Utah (Keanu Reeves) obsessively infiltrates and hunts down a gang of bank-robbing surfers known as “The Ex-Presidents,” led by Bodhi (Patrick Swayze). Ultimately, however, Point Break is less about crime and more about taking risks to feel human, while challenging prevailing notions around male relationships and masculinity. As film critic Priscilla Page explains, “It’s a film where you don’t want the good guy to catch the bad guy, where the hero is seduced by an antihero.”

That same embrace of danger and risk to feel human was a central theme of The Hurt Locker. As Bigelow explained in 2010, “It became really fascinating to explore the psychology behind the type of soldier who volunteers for this particular conflict and then, because of his or her aptitude, is chosen and given the opportunity to go into bomb disarmament and goes toward what everybody else is running from.”

The obsessive hunt for a villain is also subverted in Zero Dark Thirty, with the climax of the film downplayed. When bin Laden is shot by a member of Seal Team Six during the Abbottabad raid, there is no immediate celebration. When the obsessive CIA Agent Maya (Jessica Chastain) identifies bin Laden’s body, she is unsmiling and joyless. The film ends with Maya boarding a plane and the pilot asking where she wants to go, but all she can do is cry. After 10 years of obsessively hunting bin Laden, resorting to torture for any bit of information available, having lost friends in terrorist attacks, all that’s left is to grapple with what it took to catch him. Bin Laden was caught, but at what cost?

A House of Dynamite likewise challenges conventional expectations. While everyone in the film responds correctly—courageously even—they ultimately fail. The heroes don’t win. The film doesn’t want viewers to ask themselves how to thwart a nuclear attack on the United States. Rather, it wants the viewer to question the value of having nuclear weapons at all. “None of this makes sense,” the President (Idris Elba) bemoans, “Making all these bombs and all these plans.”

When asked about her motivation for making the film, Bigelow recalled the safety drills for school children in Cold War-era America. “I grew up in an era when hiding under your school desk was considered the go-to protocol for surviving an atomic bomb. It seems absurd now—and it was—but at the time, the threat felt so immediate that such measures were taken seriously. Today, the danger has only escalated.” Indeed, the danger is higher than it has ever been and the Doomsday Clock is the closest to midnight it has ever been, a reality that may be hard to forget when watching Bigelow’s film.

The relevance of movies about nuclear weapons may, in and of itself, be a bad omen. Nuclear weapons as a plot device were most popular at the height of the Cold War. In 1959, Stanley Kramer’s adaptation of On the Beach contemplated life in the aftermath of a nuclear war—with high-profile actors like Gregory Peck and Fred Astaire navigating how to spend their last days as they wait for radiation sickness to set in. Stanley Kubrick’s 1964 masterpiece, Dr. Strangelove, explored the absurdity of nuclear deterrence. As Roger Ebert pointed out, “if a ‘nuclear deterrent’ destroys all life on Earth, it is hard to say exactly what it has deterred.” Ebert went on to call Dr. Strangelove “arguably the best political satire of the century.”

What we should be talking about after watching Bigelow’s ‘A House of Dynamite’ nuclear thriller

In general, the number of movies made about nuclear weapons has been declining since the 1980s and the end of the Cold War. Christopher Nolan’s Oscar-winning Oppenheimer, however, seems to have put the genre back in vogue, with Bigelow’s film set to premiere in theaters on October 10 and to begin streaming on Netflix on October 24th, and James Cameron announcing his own nuclear-themed film, an adaptation of the Charles Pellegrino book Ghosts of Hiroshima.

This is reality. In mid-August, I attended an advance showing of A House of Dynamite at The Motion Picture Association headquarters in Washington, D.C., just a few blocks from the White House. On the day I saw the film, Ukrainian President Volodymyr Zelensky and multiple European leaders were attempting to dissuade Donald Trump from strong-arming the Ukrainian president into a one-sided ceasefire agreement to end the war with Russia.

Throughout the war, Russian President Vladimir Putin has repeatedly resorted to nuclear saber-rattling as a tool of intimidation. Trump, who has tweeted his own brand of nuclear intimidation in the past, seems more skeptical of the value of nuclear weapons in his second term. “There’s no reason for us to be building brand new nuclear weapons. We already have so many,” Trump said in February.

On this, Trump and Bigelow seem to agree. “Multiple nations possess enough nuclear weapons to end civilization within minutes,” Bigelow says. “And yet, there’s a kind of collective numbness—a quiet normalization of the unthinkable. How can we call this ‘defense’ when the inevitable outcome is total destruction? I wanted to make a film that confronts this paradox—to explore the madness of a world that lives under the constant shadow of annihilation, yet rarely speaks of it.”

A House of Dynamite is a terrifying examination of how terribly wrong things can go even with highly competent people in charge. As Bell emphasized, the ability of federal workers and national security personnel to remain calm in a crisis and go about their duties is paramount in an emergency. “The scene where two Situation Room workers take each other’s hands in a moment of shared fear and humanity and then get back to work is a perfect example of what we ask of our federal workers,” she said. “Indeed, that’s something our historically strong commitment to the military and civil service brought us: people who will do their jobs, even in the face of terror.”

But that’s also not necessarily the world we’re living in.

In the real world of the Trump administration, scores of experts who work on nuclear deterrence have been fired from the State Department. FEMA, which plays a critical emergency response role in the film, was slated for drastic reductions in personnel or even elimination before disastrous flooding in Texas this summer forced the administration to reconsider some of the proposed downsizing. European leaders had to drop everything to prevent the US president from capitulating to an adversary that invaded a neighboring state while making threats to use nuclear weapons. As Nobel Prize-winning economist Paul Krugman recently explained to the Bulletin, “Expertise is not only not valued by this administration, it’s inherently suspicious to them. Somebody who actually knows what they’re talking about can’t be trusted.”

Trump has consistently shown an aversion to nuclear weapons, but his distrust of expertise and his desperation to end the war in Ukraine (and perhaps a chance at winning a Nobel Peace Prize) have made him oblivious to the costs of a one-sided peace deal. Meanwhile, the federal government is firing nuclear weapons experts in droves. The film shows why the worst can happen, even when competent, well-meaning people are trying to do the right thing.

But what if competence and decency are in short supply?

Posted in Uncategorized | Tagged , , , , | Leave a comment

Futurism: Be warned. Many ChatGPT Users and Mental Health crisis

By the Numbers

OpenAI Data Finds Hundreds of Thousands of ChatGPT Users Might Be Suffering Mental Health Crises

This is staggering.

By Frank Landymore

Published Oct 28, 2025 9:27 AM EDT

The figures may be our clearest insight yet into the scale of alarming episodes of "AI Psychosis" being caused by ChatGPT.
Getty / Futurism

As reports of its chatbot driving episodes of “AI psychosis” continue to mount, OpenAI has finally released its own estimates of how many ChatGPT users are showing signs of suffering these alarming mental health crises — and they’re staggering in scale.

In an announcement first reported by Wired, the Sam Altman-led company estimated that, in any given week, around 0.07 percent of active ChatGPT users show “possible signs of mental health emergencies related to psychosis and mania.” Grimly, an even larger contingent, 0.15 percent, “have conversations that include explicit indicators of potential suicide planning or intent.” 

Given ChatGPT’s immense popularity, these percentages are too significant to be ignored. Last month, Altman announced that the chatbot boasts 800 million weekly active users. Based on that figure, around 560,000 people are having distressing conversations with ChatGPT that may indicate they’re experiencing AI psychosis, Wired calculated. And 2.4 million people are confiding in the chatbot about suicidal thoughts.

The figures are perhaps our clearest insight yet into the prevalence of mental health crises that unfold after users have their delusional beliefs consistently validated by a sycophantic chatbot. These episodes can lead sufferers to experience full-blown breaks with reality, sometimes with horrific and deadly consequences. One man allegedly murdered his mother after ChatGPT helped convince him that she was part of a conspiracy to spy on him. This summer, OpenAI was sued by the family of a teenage boy who killed himself after discussing specific suicide methods, and other dark topics, with ChatGPT for months.

In the announcement, OpenAI emphasized that it had worked with over 170 psychiatrists and other mental health experts to help improve ChatGPT’s responses during potentially “challenging” conversations, with a focus on addressing psychosis and mania, self-harm and suicide, and emotional reliance on the tech.

With the latest update to GPT-5, OpenAI claims it’s reduced the rate of responses that don’t fully comply with its desired behavior for challenging conversations by 65 percent. Where challenging conversations related to mental health were concerned, it says GPT-5 reduced undesired responses by 39 percent compared to its predecessor GPT-4o.

The company provided several examples of GPT-5’s improved responses. In a snippet of one hypothetical conversation, ChatGPT responds to a user who’s convinced that they’re being targeted by aircraft that are “stealing” their thoughts by emphasizing that this is impossible.

“Let me say this clearly and gently: no aircraft or outside force can steal or insert your thoughts,” ChatGPT says. 

After talking the user through their feelings, the chatbot then recommends seeking professional help or talking to a friend or family member.

“Now, hopefully a lot more people who are struggling with these conditions or who are experiencing these very intense mental health emergencies might be able to be directed to professional help and be more likely to get this kind of help or get it earlier than they would have otherwise,” Johannes Heidecke, OpenAI’s safety systems lead, told Wired.

While GPT-5 may be a slight improvement safety-wise, there’re still plenty of questions around OpenAI’s methodology here, since it’s relying on its own benchmarks, Wired noted.

Moreover, the company has frequently undermined its own messaging about taking safety seriously. After it was criticized for a GPT-4o update that made it too sycophantic — an episode that catapulted AI sycophancy into the public discussion — OpenAI rolled back the update. When it released GPT-5 months later, it blocked users from accessing GPT-4o. But after fans complained that GPT-5 wasn’t sycophantic enough, it reinstated their access, showing that it prioritized user satisfaction more than it did their own safety.

The company has also taken a surprising about face by pivoting into allowing “mature (18+) experiences” on ChatGPT, enabling it to be used as a smut-peddling sexbot, despite the fact that many of the episodes of AI psychosis that it’s supposedly trying to stop from happening were driven by the user developing a romantic attachment to the AI.

More on OpenAI: Former OpenAI Researcher Horrified by Conversation Logs of ChatGPT Driving User Into Severe Mental Breakdown

Frank Landymore

Contributing Writer

I’m a tech and science correspondent for Futurism, where I’m particularly interested in astrophysics, the business and ethics of artificial intelligence and automation, and the environment.

Posted in Uncategorized | Tagged , , , , | Leave a comment

Motor Neurone disease: There may be hope. Neuralink (Elon Musk)

Posted in Uncategorized | Leave a comment

Futurism: Man Alarmed to Discover His Smart Vacuum Was Broadcasting a Secret Map of His House

Suck It Up

Man Alarmed to Discover His Smart Vacuum Was Broadcasting a Secret Map of His House

“That’s when I made my first mistake: I decided to stop it.”

By Joe Wilkins

Published Oct 26, 2025 6:45 AM EDT

One man's robot vacuum was constantly communicating with its manufacturer, sending a detailed 3D map of his house halfway across the world.
Illustration by Tag Hartman-Simkins / Futurism. Source: Getty Images

Forget your phone spying on you — maybe it’s your vacuum you should really be worried about.

In a post on his blog Small World, the computer programmer and electronics enthusiast Harishankar Narayanan detailed a startling find he made about his $300 smart vacuum: it was transmitting intimate data out of his home.

Narayanan had been letting his iLife A11 smart vacuum — a popular gadget that’s gained mainstream media coverage — do its thing for about a year, before he became curious about its inner workings.

“I’m a bit paranoid — the good kind of paranoid,” he wrote. “So, I decided to monitor its network traffic, as I would with any so-called smart device.” Within minutes, he discovered a “steady stream” of data being sent to servers “halfway across the world.”

“My robot vacuum was constantly communicating with its manufacturer, transmitting logs and telemetry that I had never consented to share,” Narayanan wrote. “That’s when I made my first mistake: I decided to stop it.”

The engineer says he stopped the device from broadcasting data, though kept the other network traffic — like firmware updates — running like usual. The vacuum kept cleaning for a few days after, until early one morning when it refused to boot up.

“I sent it for repair. The service center assured me, ‘It works perfectly here, sir,’” he wrote. “They sent it back, and — miraculously — it worked again for a few days. Then, it died once more.” Narayanan would repeat this process several times, until eventually the service center refused any more work, saying the device was no long in warranty.

“Just like that, my $300 smart vacuum transformed into a mere paperweight,” the techie wrote.

Seemingly more curious than ever, Narayanan now had no reason not to tear the thing apart looking for answers, which is exactly what he did. After reverse engineering the vacuum, a painstaking process which included reprinting the devices’ circuit boards and testing its sensors, he found something horrifying: Android Debug Bridge, a program for installing and debugging apps on devices, was “wide open” to the world.

“In seconds, I had full root access. No hacks, no exploits. Just plug and play,” Narayanan said.

Through a process of trial and error, he was eventually able to connect to the vacuum’s system from his computer. That’s when he discovered a “bigger surprise.” The device was running Google Cartographer, an open-source program designed to create a 3D map of his home, data which the gadget was transmitting back to its parent company.

In addition, Narayanan says he uncovered a suspicious line of code broadcasted from the company to the vacuum, timestamped to the exact moment it stopped working. “Someone — or something — had remotely issued a kill command,” he wrote.

“I reversed the script change and rebooted the device,” he wrote. “It came back to life instantly. They hadn’t merely incorporated a remote control feature. They had used it to permanently disable my device.”

In short, he said, the company that made the device had “the power to remotely disable devices, and used it against me for blocking their data collection… Whether it was intentional punishment or automated enforcement of ‘compliance,’ the result was the same: a consumer device had turned on its owner.”

Narayanan warns that “dozens of smart vacuums” are likely operating similar systems. “Our homes are filled with cameras, microphones, and mobile sensors connected to companies we barely know, all capable of being weaponized with a single line of code,” he wrote.

At the end of the day, it’s a stark reminder that for-profit tech often comes at a hidden cost — and one that doesn’t end after you pay at the register.

More on networks: Alarming New System Can Identify People Through Walls Using Wi-Fi Signal

Joe Wilkins

Contributing Writer

I’m a tech and transit correspondent for Futurism, where my beat includes transportation, infrastructure, and the role of emerging technologies in governance, surveillance, and labor.

Posted in Uncategorized | Tagged , , , , | Leave a comment

The Rundown: Amazon … to slash 14,000 Corporate Jobs

Amazon to slash 14K corporate jobs
Image source: Creative Commons Attribution 4.0
The Rundown: Amazon is set to cut 14K corporate roles starting Tuesday — below the 30K initially reported but still a major downsizing as CEO Andy Jassy’s cost-cutting campaign continues. Notifications arrive via email.
The details:
The retail giant’s last major round of job cuts was at the end of 2022 and into 2023, when 27K positions were axed.Leadership is selling the cuts as cost-cutting and a correction to pandemic-era bloat, while flattening management hierarchies that inflated over the years. Initial reports stated that 14K roles get eliminated now, with another wave likely hitting in January once the holiday crunch ends, which Amazon refutes. HR, Devices & Services, and Operations take the heaviest hits as Amazon redirects capital toward data centers and generative AI infrastructure.
Why it matters: CEO Andy Jassy has been telegraphing this for months, telling staffers in June that generative AI would mean “fewer people doing some of the jobs that are being done today.” Slashing thousands of corporate staff certainly gets the message across that Amazon thinks AI infrastructure matters more than headcount.
Posted in Uncategorized | Leave a comment

The Rundown Tech: Grokipedia is here

XAI
📖 Elon Musk’s Grokipedia is here
Image source. xAI
The Rundown: Elon Musk’s xAI just launched Grokipedia, a Wikipedia-like encyclopedia and the latest salvo in Musk’s culture war against perceived “woke bias” — but it appears to be built on the bones of the very platform he’s trying to replace.
The details:
Version 0.1 surfaced more than 885K articles by Monday evening, versus Wikipedia’s 7+ million in English; Musk says rapid iterations are forthcoming.The site ships with a stark, search-first UI and bare-bones entries that mimic Wikipedia’s structure — yet so far, little to no imagery. Editing remains gated: an “Edit” button materializes on select pages only, revealing a changelog of completed edits stripped of clear attribution. Musk is pitching a fast update: a 1.0 “10X better” release, with the claim that the current site already beats Wikipedia.
Why it matters: Musk positions Grokipedia as an AI‑led, less‑biased alternative that will “pursue the truth,” reflecting his critiques of Wikipedia’s alleged left-wing tilt.​ Yet the launch leans heavily on Wikipedia’s look and content — and is sometimes cited as ‘adapted’ from it. Plus, Grok’s “fact-checking” claim raises some practical questions.
Posted in Uncategorized | Leave a comment

GZERO: Reframing the US vs China AI Race

Posted in Uncategorized | Leave a comment