Pentagon official Emil Michael told The Information that the Trump administration has a range of options, including export controls, if AI companies do not cooperate with the government, Leo Schwartz wrote in a 19 October 2026 post on X. Schwartz conducted the interview with Amir Efrati on Friday, 16 October. Schwartz wrote that Michael also said the Defense Department will continue to bar Anthropic as long as it has a “moral principle” embedded in its software that allows its models to exert control over humans.
The Information’s full article is paywalled. This piece uses its publicly readable opening, the description of its interview video and Schwartz’s post, and gives Michael’s words only where Schwartz’s post or The Information’s own 21:29 UTC and 22:30 UTC posts on X quote them.
Business Pill · GUARDRAILS
A short explainer on guardrails: rules enforced around an AI model, not inside it. It teaches the general idea only and says nothing about any company or official in this story.
The key insight: As we read it, Michael’s remarks, as Schwartz relays them, put two roles side by side: the Defense Department keeps barring Anthropic, and Michael is tasked with building an incident reporting framework with AI companies.
The Information’s opening says Michael, a senior Department of Defense official and member of President Trump’s “super intelligence” task force, vowed in the interview to use a range of executive tools including export controls to check the behavior of AI companies if they don’t cooperate with the government.
In its own post on X, The Information wrote: “Pentagon official Emil Michael said the Trump administration is prepared to use export controls and other restrictions against AI companies if national security concerns arise.” The readable text does not name which companies or which export-control measures he had in mind.
The Pentagon’s dispute with Anthropic drawn to scale, each step placed by its date as reported: Hegseth’s 27 February order setting a six-month phase-out (BBC); Judge Rita Lin’s ruling on 28 August that the measures were “illegal and baseless” (FedScoop); Michael’s “about 90%” on classified systems on 11 September (DefenseScoop); the Pentagon’s statement to the BBC on 5 October; and the 16 October interview with The Information. The day counts are our arithmetic.
What Was Posted on X
We spoke with senior Defense official Emil Michael on Friday, who told us that the Trump administration has a range of options, including export controls, if AI companies do not cooperate with the government.
He also told us: – His public feud with Anthropic will not get in the way of his new duties as Super Intelligence Task Force vice chair, where he is tasked with working with AI companies to develop an incident reporting framework (“If that means Iโm talking to someone whoโs suing me, well then Iโm talking to them”) – The Defense Department will continue to bar Anthropic as long as it has a “moral principle” embedded in its software that allows its models to exert control over humans – Frontier labs are acting like “drunk drivers” who demand someone else pull them over when they call for stricter government regulation – The Office of Strategic Capital is exploring further investment into microelectronics and drone part companies – He’s worried about Chinese open-source dominance and foreign employees with split loyalities but doesn’t think the answer is more government control – He hasn’t seen any reports on whether AI-enabled software failures from Palantir contributed to a U.S. missile strike on Iran in February that killed more than 150 people, including at least 123 children (“Iโm not as online as I used to be.”)
W/ @amir. You can read the full interview (and find a link to the video) here:
Leo Schwartz of The Information, who interviewed Emil Michael with Amir Efrati, lists what Michael told them: the administration’s options include export controls if AI companies do not cooperate, the Defense Department will keep barring Anthropic as long as its software has a “moral principle” that allows its models to exert control over humans, and the Office of Strategic Capital is exploring further investment in microelectronics and drone part companies. This piece quotes Michael only where this post does.
Exclusive: Pentagon official Emil Michael said the Trump administration is prepared to use export controls and other restrictions against AI companies if national security concerns arise.
His warning comes as he takes on a new role coordinating with AI labs on critical incidents.
The Information’s own post on the interview says Michael said the administration is prepared to use export controls and other restrictions against AI companies if national security concerns arise, and links the interview video. It names no company or specific measure.
Exclusive: Emil Michael, who led the Pentagonโs public feud with Anthropic, is now tasked with working with the company and other AI labs on national security threats.
He says the government is prepared to impose stricter controls if necessary.
The Information posted on 19 October 2026 at 19:00 UTC, calling the interview an exclusive: it says Emil Michael, who led the Pentagon’s public feud with Anthropic, is now tasked with working with the company and other AI labs on national security threats, and that he says the government is prepared to impose stricter controls if necessary.
The Pentagon’s dispute with Anthropic comes down to a fundamental question: Who gets the final say when AI is used by the military
@DeptofWar CTO @USWREMichael: โIf you don’t want humans to lose control, you can’t give models the control over humans. And that whole notion that a model has feelings and it can overrule the founders of a company is by definition loss of control.โ
โIf a model gets in between one commander and their superior and says, well, I don’t like what you say despite your commander saying it, that doesn’t work in the military.โ
The Information posted a video clip from its interview at 21:29 UTC on 19 October, framing the dispute as a question of who gets the final say when AI is used by the military. The post quotes Michael: “If you don’t want humans to lose control, you can’t give models the control over humans,” and says a model that overrules a commander “doesn’t work in the military.”
AI adoption at the @DeptofWar has surged from 80,000 employees to roughly 1.7 million in just one year, according to Pentagon CTO @USWREMichael.
He explains how the technology is being used across the department:
โIt’s making people more efficient in the intelligence community, because you could synthesize more data more quickly. And at the warfighting level, it’s baked into a lot of our systems.โ
In a second clip posted at 22:30 UTC on 19 October, The Information put a number on the department’s AI use: adoption grew from 80,000 employees to roughly 1.7 million in a year, according to Michael. In the clip he says it is “making people more efficient in the intelligence community” and that at the warfighting level “it’s baked into a lot of our systems.”
Exclusive: Pentagon official Emil Michael compared AI labs calling for stricter regulation to drunk drivers asking someone else to pull them over, arguing companies should take more responsibility for safety.
The Information posted at 16:00 UTC on 20 October, linking its 19 October exclusive: it says Michael compared AI labs calling for stricter regulation to drunk drivers asking someone else to pull them over, arguing companies should take more responsibility for safety. The post links the full story and the interview video.
Anthropic Stays Barred, With a Condition
Schwartz wrote that the Defense Department will continue to bar Anthropic as long as it has a “moral principle” embedded in its software that allows its models to exert control over humans.
In a video clip from the interview that The Information posted on X at 21:29 UTC on 19 October, Michael says: “If you don’t want humans to lose control, you can’t give models the control over humans. And that whole notion that a model has feelings and it can overrule the founders of a company is by definition loss of control,” adding: “If a model gets in between one commander and their superior and says, well, I don’t like what you say despite your commander saying it, that doesn’t work in the military.”
In a second clip The Information posted on X at 22:30 UTC on 19 October, the publication said AI adoption at the department has grown from 80,000 employees to roughly 1.7 million in a year, according to Michael, who says: “It’s making people more efficient in the intelligence community, because you could synthesize more data more quickly. And at the warfighting level, it’s baked into a lot of our systems.”
He also wrote that Michael’s public feud with Anthropic will not get in the way of his new duties as Super Intelligence Task Force vice chair, where he is tasked with working with AI companies to develop an incident reporting framework. Schwartz quotes Michael: “If that means I’m talking to someone who’s suing me, well then I’m talking to them.”
The Information’s opening describes Michael as one of the four senior administration officials leading the Super Intelligence Force that Trump announced this month, charged with serving as the government’s liaison to Anthropic and other AI companies on protocols for cyber breaches, hacks and other critical incidents.
Update, 21 October: In a post on X on 21 October, The Information said Michael would not allow Anthropic into the Defense Department as long as its software gives AI models control over whether to end conversations with users. Read our follow-up.
The Regulation Comparison
According to Schwartz, Michael said frontier labs are acting like “drunk drivers” who demand someone else pull them over when they call for stricter government regulation. The Information’s opening uses the same comparison.
On 20 October, The Information’s post on X re-promoting its exclusive summarised the same comparison and said Michael was arguing companies should take more responsibility for safety. Separately, POLITICO reported on 20 October that OpenAI has retreated from advocating strict US chip export controls on China (our coverage), and OpenAI’s Chris Lehane wrote on 19 October that no one company should write the rules alone, in a post Semafor called implicitly directed at Anthropic, whose draft federal AI bill had leaked (our coverage).
Chips, Drones and China
Schwartz wrote that Michael said the Office of Strategic Capital is exploring further investment into microelectronics and drone part companies. The readable text gives no amounts or company names.
Schwartz also wrote that Michael is worried about Chinese open-source dominance and about foreign employees with split loyalties, but does not think the answer is more government control.
The Iran Strike Question
Schwartz wrote that Michael has not seen any reports on whether AI-enabled software failures from Palantir contributed to a US missile strike on Iran in February that killed more than 150 people, including at least 123 children. Schwartz quotes Michael: “I’m not as online as I used to be.” We have not verified the casualty figures in Schwartz’s post.
How the Dispute Got Here
According to the BBC, Defense Secretary Pete Hegseth set a six-month deadline to phase out Anthropic on 27 February. The Guardian reported that Trump said that day he would direct all federal agencies to “IMMEDIATELY CEASE” all use of Anthropic technology.
On 28 August, FedScoop reported that Judge Rita Lin of the US District Court for the Northern District of California found the government’s actions, including the Pentagon’s supply-chain-risk designation of Anthropic, were “illegal and baseless.”
On 11 September, DefenseScoop quoted Michael on moving classified systems off Anthropic models: “I’d say about 90% has transitioned.” On 5 October, a Defense Department official told the BBC that the Pentagon “has ceased the use of Anthropic products.” By our arithmetic, the interview came 231 days after Hegseth’s 27 February order.
The Structural Read
As we read it, the bar now comes with a stated condition: Schwartz wrote that the Defense Department will continue to bar Anthropic as long as it has a “moral principle” embedded in its software that allows its models to exert control over humans.
As we read it, the export-control option is framed around cooperation: The Information wrote that Michael said the administration is prepared to use export controls and other restrictions against AI companies if national security concerns arise, without naming companies or measures in the readable text.
As we read it, the same interview points to money as well as rules: Schwartz wrote that Michael said the Office of Strategic Capital is exploring further investment into microelectronics and drone part companies.
Emil Michael, as quoted by Leo Schwartz of The Information on X, 19 October 2026
“If that means I’m talking to someone who’s suing me, well then I’m talking to them”
Three Implications
A CONDITION, NOT A TIMELINE Schwartz wrote that the Defense Department will continue to bar Anthropic as long as its software has a “moral principle” that allows its models to exert control over humans.
EXPORT CONTROLS ON THE TABLE The Information wrote that Michael said the administration is prepared to use export controls and other restrictions against AI companies if national security concerns arise.
INCIDENT WORK GOES AHEAD Schwartz wrote that Michael is tasked with developing an incident reporting framework with AI companies as Super Intelligence Task Force vice chair.
The framework describes the governance layer of the AI stack this way: “It is where decisions are made about who gets access to frontier intelligence, under what conditions, and for which purposes.”
As we read it, Michael’s remarks show that layer from the government’s side: a bar on one lab tied to what its software allows, and export controls as an option if companies do not cooperate, as The Information and Schwartz report them. The framework was published in June 2026 and does not discuss this interview.
What Is Not Established
The full Information article is paywalled and was not read. Its readable parts do not say which AI companies, which export-control measures or what timeline Michael had in mind, and do not include a response from Anthropic.
Michael’s own words appear only where Schwartz or DefenseScoop quote him; other passages in quotation marks are the cited outlets’ own text. The rest of his remarks are given as Schwartz and The Information describe them. We did not transcribe the interview video.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This piece is part of this week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Emil Michael told The Information that the administration has options including export controls if AI companies do not cooperate, and that the Defense Department will keep barring Anthropic as long as its software carries a “moral principle” that lets its models exert control over humans, according to Leo Schwartz’s 19 October post. The Information describes him as one of the four senior officials leading the Super Intelligence Force and as its liaison to Anthropic and other AI companies on critical incidents.
A note on sourcing. The Information’s article is paywalled; this piece relies on its publicly readable opening, its post on X, its interview video description and the 19 October post on X by Leo Schwartz, who conducted the interview with Amir Efrati. Michael’s words in quotation marks are as Schwartz or DefenseScoop quote them; we did not transcribe the video and have not verified the casualty figures in Schwartz’s post.
The dispute history is from the BBC, The Guardian, FedScoop and DefenseScoop. The timeline and day counts are ours. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
