AI misidentification sparks US interception
6m
An AI-generated U.S. intelligence report wrongly identified a Chinese vessel in the Middle East as carrying nuclear components, prompting U.S. forces to plan an interception, officials said.
Now imagine that during a crisis an AI draws from a huge amount of data to conclude that Russia or China are preparing a nuclear first strike and an under-pressure analyst sends that immediately up the chain.
The US military almost boarded a Chinese cargo ship because an intelligence analyst trusted a false AI-generated conclusion that the ship was carrying components for a nuclear-weapons program. edition.cnn.com/2026/09/18/p...
September 19, 2026 at 7:19 AM
Now imagine that during a crisis an AI draws from a huge amount of data to conclude that Russia or China are preparing a nuclear first strike and an under-pressure analyst sends that immediately up the chain.
HUGE EXCLUSIVE from @bloomberg.com
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators have discovered that flawed intelligence, outdated imagery and an overreliance on AI contributed to a missile strike that killed 123 children in Minab.
www.bloomberg.com
September 18, 2026 at 9:24 PM
HUGE EXCLUSIVE from @bloomberg.com
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Reposted by Suresh Venkatasubramanian, Graham Webster
Hallucinating AI "almost started a war" by FALSELY claiming China was smuggling nuclear materials.
"The US military swung into action with plans to intercept the vessel... Military planes were in the air"
"The US military swung into action with plans to intercept the vessel... Military planes were in the air"
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
edition.cnn.com
September 18, 2026 at 7:22 PM
Hallucinating AI "almost started a war" by FALSELY claiming China was smuggling nuclear materials.
"The US military swung into action with plans to intercept the vessel... Military planes were in the air"
"The US military swung into action with plans to intercept the vessel... Military planes were in the air"
The US military almost boarded a Chinese cargo ship because an intelligence analyst trusted a false AI-generated conclusion that the ship was carrying components for a nuclear-weapons program. edition.cnn.com/2026/09/18/p...
September 19, 2026 at 6:47 AM
The US military almost boarded a Chinese cargo ship because an intelligence analyst trusted a false AI-generated conclusion that the ship was carrying components for a nuclear-weapons program. edition.cnn.com/2026/09/18/p...
CNN: "US military had close call after using AI for false intelligence report [claiming that a] Chinese ship in the Middle East was transporting components of a nuclear weapons program ... armed members of the US military were preparing to board the ship."
edition.cnn.com/2026/09/18/p...
edition.cnn.com/2026/09/18/p...
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
edition.cnn.com
September 19, 2026 at 5:42 AM
CNN: "US military had close call after using AI for false intelligence report [claiming that a] Chinese ship in the Middle East was transporting components of a nuclear weapons program ... armed members of the US military were preparing to board the ship."
edition.cnn.com/2026/09/18/p...
edition.cnn.com/2026/09/18/p...
This is horrifying.
Please join me in calling for an immediate return to getting our false intelligence about WMDs from the CIA and other organic sources.
Please join me in calling for an immediate return to getting our false intelligence about WMDs from the CIA and other organic sources.
Absolutely horrific: Thanks to a false intelligence "generated with the help of artificial intelligence," the US military nearly carried out an operation aboard a Chinese ship because the incorrect intelligence said it was carrying nukes, according to @cnn.com. edition.cnn.com/2026/09/18/p...
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
edition.cnn.com
September 18, 2026 at 6:39 PM
This is horrifying.
Please join me in calling for an immediate return to getting our false intelligence about WMDs from the CIA and other organic sources.
Please join me in calling for an immediate return to getting our false intelligence about WMDs from the CIA and other organic sources.
Makes me think of Stanislav Petrov, one of the greatest heroes in history. We're going to need to build systems that don't rely on people like him to stop a rush to judgment.
Lot of issues here but one part that sticks out in light of the almost-attacked-a-Chinese-ship story today: people with very important jobs where lives are at stake have bought the hype and outsourced their whole brains to AI. We're going to see more and more of this.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
September 19, 2026 at 12:55 AM
Makes me think of Stanislav Petrov, one of the greatest heroes in history. We're going to need to build systems that don't rely on people like him to stop a rush to judgment.
it's really essential to underline that the use of failing or problematic software tools/AI tech IN NO WAY ABSOLVES THE ACTORS OF CRIMINAL CULPABILITY.
Lot of issues here but one part that sticks out in light of the almost-attacked-a-Chinese-ship story today: people with very important jobs where lives are at stake have bought the hype and outsourced their whole brains to AI. We're going to see more and more of this.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
September 18, 2026 at 11:44 PM
it's really essential to underline that the use of failing or problematic software tools/AI tech IN NO WAY ABSOLVES THE ACTORS OF CRIMINAL CULPABILITY.
Wait, "a shipment of F-35 parts bound for the United States for repairs was rerouted as it transited the Pacific Ocean from Australia" to... Hong Kong?! That's like flying from Chicago to NYC but landing in Los Angeles instead. www.politico.com/news/2026/09...
Hill investigates after F-35 parts diverted to Hong Kong
The incident raises the possibility that top secret weapons technology may have been exposed to a foreign adversary — including China.
www.politico.com
September 18, 2026 at 7:44 PM
Wait, "a shipment of F-35 parts bound for the United States for repairs was rerouted as it transited the Pacific Ocean from Australia" to... Hong Kong?! That's like flying from Chicago to NYC but landing in Los Angeles instead. www.politico.com/news/2026/09...
Fuck these motherfuckers and fuck their motherfucking machines.
HUGE EXCLUSIVE from @bloomberg.com
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators discovered that flawed intel, outdated imagery & overreliance on AI contributed to a missile strike that killed 123 children in Minab
🎁 www.bloomberg.com/graphics/202...
Inside the US ‘Kill Chain’ That Destroyed an Iranian School
Pentagon investigators have discovered that flawed intelligence, outdated imagery and an overreliance on AI contributed to a missile strike that killed 123 children in Minab.
www.bloomberg.com
September 18, 2026 at 10:11 PM
Fuck these motherfuckers and fuck their motherfucking machines.
*stares in Stanislav Petrov*
Sources: the US "almost started a war" after an AI-assisted report hallucinated that a Chinese ship was carrying nuclear weapons components in the Middle East (CNN)
Main Link | Techmeme Permalink
Main Link | Techmeme Permalink
September 18, 2026 at 6:03 PM
*stares in Stanislav Petrov*
by Ryan Goodman Reposted by Steven Heydemann
"The extraordinary move to punish the Royal Canadian Mounted Police ... was specific to the F.B.I. Other parts of the U.S. government involved in intelligence-sharing continued with their operations."
www.nytimes.com/2026/09/18/u...
www.nytimes.com/2026/09/18/u...
F.B.I. Temporarily Halted Cooperation With Canada Over References to 9/11
The abrupt move by Kash Patel, the bureau’s director, came as President Trump was ramping up his attacks on the country.
www.nytimes.com
September 18, 2026 at 5:48 PM
"The extraordinary move to punish the Royal Canadian Mounted Police ... was specific to the F.B.I. Other parts of the U.S. government involved in intelligence-sharing continued with their operations."
www.nytimes.com/2026/09/18/u...
www.nytimes.com/2026/09/18/u...
Reposted by Bessma Momani
Exclusive: An AI-assisted intel report, circulated across the US military during the Iran war, nearly triggered an operation but was false, sources say. https://cnn.it/4y9MBJO
September 18, 2026 at 5:31 PM
Exclusive: An AI-assisted intel report, circulated across the US military during the Iran war, nearly triggered an operation but was false, sources say. https://cnn.it/4y9MBJO
So how close are we, really, to an AI generated global conflagration? Probably a lot closer then we once imagined.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
www.cnn.com
September 18, 2026 at 7:19 PM
So how close are we, really, to an AI generated global conflagration? Probably a lot closer then we once imagined.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
Reposted by Brendan Nyhan, Stephen A. Meserve, David H. Kaye
Lot of issues here but one part that sticks out in light of the almost-attacked-a-Chinese-ship story today: people with very important jobs where lives are at stake have bought the hype and outsourced their whole brains to AI. We're going to see more and more of this.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
September 18, 2026 at 10:29 PM
Lot of issues here but one part that sticks out in light of the almost-attacked-a-Chinese-ship story today: people with very important jobs where lives are at stake have bought the hype and outsourced their whole brains to AI. We're going to see more and more of this.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
by Nathan P. Kalmoe Reposted by David R. Miller
This episode is not just embarrassing for the military -- it shows for the millionth time the self-destructive insanity of leaders in every institution on earth pumping out pro-AI propaganda, from business to edu to news to government to unions. AI has no place being trusted for anything important.
It appears the analyst in question didn’t even bother to check his own work or the conclusions reached by the chatbot
September 18, 2026 at 5:52 PM
This episode is not just embarrassing for the military -- it shows for the millionth time the self-destructive insanity of leaders in every institution on earth pumping out pro-AI propaganda, from business to edu to news to government to unions. AI has no place being trusted for anything important.
“AI allows you to get to a bad idea faster” www.cnn.com/2026/09/18/p...
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
www.cnn.com
September 18, 2026 at 6:56 PM
“AI allows you to get to a bad idea faster” www.cnn.com/2026/09/18/p...
This is my absolute bone core experience (with many students, told AI is the best all the way up to the uni president). *This* is what AI is doing to people in society even worse than doing a iffy job at times, just letting it make decisions
(accidentally QTed a reply the first time!)
(accidentally QTed a reply the first time!)
Lot of issues here but one part that sticks out in light of the almost-attacked-a-Chinese-ship story today: people with very important jobs where lives are at stake have bought the hype and outsourced their whole brains to AI. We're going to see more and more of this.
www.cnn.com/2026/09/18/p...
www.cnn.com/2026/09/18/p...
September 19, 2026 at 2:11 AM
This is my absolute bone core experience (with many students, told AI is the best all the way up to the uni president). *This* is what AI is doing to people in society even worse than doing a iffy job at times, just letting it make decisions
(accidentally QTed a reply the first time!)
(accidentally QTed a reply the first time!)
““AI allows you to get to a bad idea faster,” one of the sources said.”
Bad enough if it is a student writing a term paper. Much worse when it is a military “analyst” who impacts targeting.
Also, querying a chatbot does not an analyst make.
Bad enough if it is a student writing a term paper. Much worse when it is a military “analyst” who impacts targeting.
Also, querying a chatbot does not an analyst make.
Absolutely horrific: Thanks to a false intelligence "generated with the help of artificial intelligence," the US military nearly carried out an operation aboard a Chinese ship because the incorrect intelligence said it was carrying nukes, according to @cnn.com. edition.cnn.com/2026/09/18/p...
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
edition.cnn.com
September 18, 2026 at 6:58 PM
““AI allows you to get to a bad idea faster,” one of the sources said.”
Bad enough if it is a student writing a term paper. Much worse when it is a military “analyst” who impacts targeting.
Also, querying a chatbot does not an analyst make.
Bad enough if it is a student writing a term paper. Much worse when it is a military “analyst” who impacts targeting.
Also, querying a chatbot does not an analyst make.