From Washington: The Debate Over Militarizing Artificial Intelligence and the Future of War

In 2024, a truck crashed into Canaw in Moresque, where I work.

146 of our dogs needed homes fast.

We asked for help on Facebook.

Our stories spread through WhatsApp messages and Instagram reposts.

Immediately, people stepped up.

And just six hours later, every dog was fostered.

I'll never forget how our community showed up for us.

Learn how over 3.5 billion people connect to what matters with meta

at meta.com slash community.

That's pure automotive joy.

I'm Peter, the owner of Muscle Car Junior.

It started as a hobby, then I started posting about it.

Before I knew it, I built a business for storing muscle cars

on Facebook Marketplace, and the community of car lovers on Instagram.

Today, new customers send me what's app messages from all over.

Not bad for a hobby.

Learn how meta helps over 3.5 million American businesses,

like Peter's Grow, at meta.com slash community.

Sunday, March 8, 2026, I'm Jessica Rosenthal.

Just before the U.S. struck Iran last Saturday,

the Department of War said anthropic AI was out

over the company's refusal to allow broader access.

Open AI now has a new contract with the Pentagon.

It's important that we have the technology.

It's just a matter that we don't want that technology

to get out of hand.

And I think that cloud and anthropic was right.

We don't want this to be used for autonomous killing systems

that don't have a human in the loop.

I'm Dave Anthony.

Did you know, before the U.S. and Iran were foes, we were friends.

My greatest hope is that we could develop a situation

again where the U.S. and Iran are allies,

but again, was the real history of U.S. and Iran

until this unfortunate detour

that the materialist enhancing mullahs

have been taking us on for 47 years.

This is the Fox News rundown from Washington.

Just before U.S. strikes on Iran,

things came to a head between anthropic

and the Department of War.

Secretary P. Tagseth told the AI company

their restrictions hindered national security operations

that he would not allow a company to dictate the terms

under which the Pentagon makes operational decisions.

Anthropics CEO Dario Amadi told CBS News March 1st.

We believe in defeating our autocratic adversaries.

We believe in defending America.

The red lines we have drawn, we drew because we believe

that crossing those red lines is contrary to American values.

And we wanted to stand up for American values.

Those red lines were use of their systems

in autonomous weapon systems and for mass surveillance.

The Pentagon then terminated the $200 million contract

with anthropic and labeled them a supply chain risk.

Anthropics says they will challenge that designation in court.

Open AI now has a contract with the war department,

which the company says also contains red lines.

When it comes to using AI in autonomous weapon systems,

they write this in a blog post.

Their systems will not be used to independently

direct autonomous weapon systems in any case,

where law, regulation, or department policy

requires human control.

Nor will it be used to assume other high stakes decisions

that require approval by a human decision maker.

It's phasing out because it resides inside a classified system

that is key to the planners.

Retired Army Lieutenant Colonel Robert McGuinness.

Having been a strategist in the planner

in the Pentagon for decades, I understand that, you know,

you don't want rapid change on something

that you've been depending upon and putting together,

you know, all the details for months and months,

all of a sudden the rug pull out from underneath you.

So I don't think Sam Altman, Open AI,

are quite ready to replace what anthropic and clawed

have been doing for the planners in the Pentagon.

However, it's important that we understand

that the decision cycle has been radically improved

for our benefit.

And we're seeing that played out not only in Iran,

but elsewhere across the Defense Department.

And so this is, it's important that we have the technology.

It's just a matter that we don't want that technology,

I think, to get out of hand.

And I think that clawed and anthropic was right.

We don't want this to be used for autonomous killing systems

that don't have a human in the loop.

And in fact, Defense Department Directive

is very specific that a human must be in the loop.

And I think that that's something that a humane,

a liberty-loving country ought to embrace.

And the other part, of course.

And I think this is where we need to look

at what President Trump has mandated,

is that anthropic is out across the U.S. government.

Well, the issue that anthropic brings up

about using this to violate our Fourth Amendment rights,

privacy, I think, is a significant issue

when you consider the data that our systems now collect

and how AI can just, you know,

piece all those together very rapidly

and perhaps make us more vulnerable

to manipulation in the future.

So these are issues of grave consequence

that I think Congress and the administration

need to work through.

We've seen the head of anthropic, Dario Amadi,

expressed some pretty serious concerns

in recent weeks to your point about AI

likening it to growing something rather than building it.

He writes about how little they still know

about what its capabilities are

that today frontier AI systems are simply not reliable enough

to power fully autonomous weapons

and that we will not knowingly provide a product

that puts America's war fighters and civilians at risk.

That at risk note.

How important is that?

Well, at risk is very important now.

We don't want to jeopardize the lives

of the young men and women that serve

in the Armed Forces United States.

I would argue, though, that we've seen

on the battlefields of Ukraine a rampant increase

in effectiveness of AI guided drones.

And I think we're seeing some of that in Iran as well.

So, you know, the technology is really speeding up far more

than many of us think.

We've gone past the generative AI to the agentic,

which is really proliferating

across defense department and elsewhere.

And there are things right on the horizon

that are just going to be mind-blowing

in comparison to what we've seen here in the recent past.

Yeah, about Ukraine.

You know, Amadi acknowledges that autonomous drones

are being used in Ukraine just as you note

that they may be very important to national defense

an autonomous drone program to a certain degree.

If open AI allows its model to be used here

for drone warfare,

what are some of the considerations,

including what could go wrong?

Well, the problem with drones is if they have an electronic trail

or connection with ground-based malware,

it could be put in, they can be manipulated,

they can be reversed.

If we know the various EW issues

and the algorithms well enough,

it's just not totally safe.

Now, if they're going to be on AI

and they're going to be autonomous totally

from any callback type of mechanism,

then that's a different proposition.

But for the most part, we don't want that.

And keep in mind also that these algorithms

are really designed to be used by the likes of the Chinese

to control hundreds of drones simultaneously

into individual, individually target them

on whether it's a tank formation

or it's a building and you want to go in certain windows

into certain offices and not explode

until they hit a particular type of material.

So the complexity of that is something far beyond

most of war fighters think about.

To your point about China,

Secretary Hegseth has said,

we need full access to AI's capabilities.

It does make you think about this conversation

we've been having now for years.

Do you think our adversaries don't have full access?

Of course they do.

And I think Hegseth's point is we need that too.

Is there something powerful to that notion

that in the end this is about a competition?

And while you can be clear-eyed about the things

that can go wrong,

that maybe shouldn't stop you from full implementation even now?

Yeah, the competition's undeniable.

The Chinese investment and artificial intelligence

they fuse everything across their civilian

and their military arena.

And so they're not going to put guardrails

in their AI military systems like we would

because they represent, from my opinion,

tyranny where we represent liberty.

And that distinction needs to be,

I think, maintained as best we can.

Now, what Pete Hegseth is talking about,

clearly, he doesn't want his hands tied.

But sometimes in a republic which is governed

by civilians, the military's hands need to be double-checked

and some cases tied.

That doesn't mean that they're not going to have

great thoughts ability.

But once again, there has to be a human in the loop.

And I want that human to be accountable

because AI is never going to be accountable.

You know, I want somebody to say,

okay, yeah, we're going to go kill that person

or thing or whatever.

I want to make sure that there's a human

with judgment that has a moral basis

that represents our republic

and not a AI entity that is objective

and only focused on, you know,

efficiency and not on morality.

Even with a human in the loop, sir,

how concerned are you given your knowledge and expertise

that an AI system, and maybe this is naive of me to say,

goes rogue or something within the AI system

happens that the human in the loop

can't control in real time.

Is that a fear that these directives address?

Yeah, well, that's what we try to do.

We try to disable our enemy's AI

with a malware or some sort of EW.

And we've, there are some success in that regard.

Now, they become more sophisticated in time

and we're trying to prevent that from happening.

But that's a real problem.

Even though we may build in the algorithm

a human in the loop at a particular decision point,

it may be that that's jeopardized totally

because of the enemy's counters.

Where you have an offensive weapon,

you also have to have a defensive weapon.

And in the AI realm, it's AI against AI.

And so it's, which has got the right twist

or the right type of information

to defeat that adversary in its own game.

And this is going to happen in microseconds.

And that's the scary thing for us humans

because we really want to have more say

than unfortunately, I'm not sure we're going to in some cases.

It does make me think, and maybe this is a bit more philosophical,

that if AI allows the Department of War

or others to conduct mass surveillance

or to use autonomous weapons,

and that effort is in large measure

to compete with an adversary,

then are we saying we have to potentially sacrifice

some civil liberties to some degree?

And if we do, are we saying

that our adversaries way of thinking wins?

Well, that's a fair question.

I would hope that we don't.

But in today's world, where

seconds matter in this business

and decisions about who are what to attack.

And if we're going to lose a city

because we were too slow to react

to a legitimate attack,

whether it's by a major weapon system

that's controlled by an AI of an adversary

and we know about it.

And we just have the next 10 seconds

to finally decide to unleash our own.

Those moral questions are very, very tough

and they need to be discussed

at the highest level of the U.S. government.

The laws and the regulations need to be put in place

because they just don't exist.

That's why I wrote what I did

because I don't see that.

I think we're of a 1990s vintage thinking

where this is the new internet

and we're going to be able to

kind of live with it for a while,

find out the glitches and then fix them in time.

AI is incredibly different.

It's very dangerous if we allow it

to go rogue and go its own direction.

And unfortunately a lot of people don't appreciate that.

And people on the hill,

you know, they're just reading the papers

like most everyone.

They don't know what the truth is

and they're not technologists

so they don't know what to believe.

You write in this foxnews.com op-ed.

Artificial intelligence has moved

from the lab to the kill chain.

What are you seeing as evidence about that

and is it just in Ukraine

or are we seeing evidence of this

in the kill chain as we strike Iran?

Well, we used the drones

and they were AI empowered or directed

and we kill people.

I don't know what the

battle damage assessment is

from those strikes

but I'm sure that the agency knows.

You know, we have overhead surveillance.

We have other means of surveilling

and we wouldn't have done that

unless we had confidence

that the systems would deliver the effects

that we were looking for.

And I assume that, you know,

the president has,

is privy to what those effects are.

So far it looks as if

everything that we've shot

has had some good measure of success.

One more finally for you.

A few years ago we were talking a lot

about a munitions shortage in this country.

Have we solved for that?

Are these strikes evidence of that

or is that sort of separate from what we're using?

Well, before I retired

from the Pentagon a year ago,

I was in that business.

So I have to be careful what I say.

We drained a lot of our

forward-placed arsenals

in order to equip the Ukrainians

with sufficient munitions

of a certain nature

to be effective against the Russians.

Most of those

most more sophisticated systems

were based upon factories

that are pretty high-tech

and you just don't stand them up overnight.

So I'm not as optimistic

that we've been able to replenish

all of those munitions

that we're firing off

a million-dollar munitions

that, you know,

a thousand-dollar drones

is not a good use of our resources.

So we have to be much smarter.

And I know we're thinking about all these things.

Whether or not

we, in fact,

have jeopardized

the stockage of arsenals,

especially when we have

potential adversaries like China

and the threat

of them taking down Taiwan

and are promised to go

to their rescue.

Those are all fair questions

and I would probably defer

to Pete Hegseth

or General Cain.

They need to answer those tough questions.

I know they're double-timing

with industry to

ratchet up the production.

But those factories,

you just can't stand those factories

up overnight.

I mean, they take years to stand up

a factory that makes

Patriot missiles,

or, you know, some of the

anti-tank systems.

They are very sophisticated

and the workers,

they don't sit around

in the local community where those factories

once were,

this is ludicrous.

So we need to understand

that, you know, it must

like the strategy

that the Pentagon put together

at the end of the Biden administration

regarding our strategic

arsenals and so forth.

We need to get serious

about that.

Otherwise, we're going to be

in a very bad situation in the future.

Retired Lieutenant Colonel

today, working hard,

planning ahead, doing everything

I can to protect them.

And it hit me.

I also need a plan for the times

where I can't be there.

That's why I'm telling you about

fabric by Gerber Life.

It was built for busy parents

who need term life insurance quickly

and easily.

Without the hassle,

if you've been putting it off,

you're not alone.

You can apply online

on your schedule right

from your couch.

You could be covered

in under 10 minutes with no

health exam required.

If you've got kids, especially

if you're young and healthy,

now's the time to lock

in low rates.

And even if you have coverage

through work, it may not be enough.

And it might not follow you

if you leave your job.

That's meatfabric.com

Slash rundown.

M-E-E-T fabric.com

Slash rundown.

Policies issued by western

southern life insurance company.

Not available in certain states.

Price is subject to underwriting

and health questions.

This US conflict with Iran

is just the latest chapter

of the long saga.

A decade ago, President Obama

took a much different approach.

America must lead not just

with our might,

but with our principles.

It shows we are stronger,

not when we are alone,

but when we bring the world together.

He hailed a joint comprehensive agreement

between the US and our allies

and Iran,

aimed at curbing Iran's nuclear program.

But President Trump called it a terrible deal

back to 1979.

When I had holochomani,

led a revolution,

and Americans were kidnapped

at the US Embassy in Tehran.

President Carter called that terrorism

in a unifying moment for the US.

We stand today as one people.

We are dedicated to the principles

and the honor of our nation.

But the Iran hostage crisis

kept going and going

and Carter lost the 1980 election

to Ronald Reagan.

With 1981, the new president

announced at an inaugural ball.

The plans have landed now, cheers.

Mountsing in the 52 hostages,

Reagan called prisoners

of war had been set free.

You can imagine their happiness.

They're preparing to board

the American planes

for the last leg of the trip.

But the first US clash with Iran

took place

way before that.

102 years ago,

the first US president

to visit what is now called Iran.

Tevi Troy is a presidential historian

who worked for a president.

As a senior White House

aide at George W. Bush,

he's also a senior fellow

at the Ronald Reagan Institute.

And he joined me earlier this week

on the Fox News rundown.

He went not necessarily for Iran itself

but for a meeting

with the so-called big three of World War II

Roosevelt, Winston Churchill, Joseph Stalin,

the young Shah, the son of Reza Shah.

And he thanked him for his hospitality.

There's even a letter in the archives

of very gracious note that he sent him.

And there was a sense that the US

and Iran were friendly nations.

And so that was a good thing.

And Truman followed up

met with him as well.

He even hosted Reza Shah Palevi

through the young Shah

in Washington, the first president

to host an Iranian later in Washington.

And then Eisenhower,

was Iran?

Well, Eisenhower, I would say,

has positive relations with Iran.

But there was a coup

where a guy named Muhammad Musadah

was the anti-Western prime minister.

And he tried to nationalize the oil industry,

which really angered the British.

The British wanted to take military action against this.

Eisenhower was against the idea.

But Musadah didn't play his hand very well.

There were large anti-Musad demonstrations.

And he might have fallen anyway.

There's a sense. And there's some historical dispute

over this that the CIA gave him a push.

And he was out and still to this day,

people who don't like America,

grumble about America's role

in that 1953 coup.

But it's not clear that it was all the US's fault

or that the US actually made it happen.

Then Eisenhower does still have good relations

with the Shah who returns to Iran

after those troubles of 1953.

And he visits Iran in 1959

and highlights the friendly relationship we have with them.

All right, President Kennedy.

He didn't go to Iran, did he?

He didn't, but the Shah's best years.

I think we're in the 1960s.

He was seen as a close American ally

and he visited America five times

during the Kennedy and Johnson years.

And he went to see the US space program at Cape Canaveral.

He got a ticker tape parade in New York City.

He had a steak dinner.

A ticker tape parade?

A ticker tape parade for the leader of Iran.

I don't really do those anymore.

Yeah, that is five.

I'm all of a certainly having not gotten a ticker tape parade

in New York City.

Nor will they anytime soon.

And then we get into the 70s.

President Nixon, he was a,

he was a big Shah backer, wasn't he?

Oh, huge Shah backer.

In fact, dating back to the 1950s

when he was vice president under Eisenhower.

So he's very positive about the Shah.

His famous national security advisor

and his later secretary of state, Henry Kissinger.

He calls the Shah an unconditional ally.

And he really sees the Shah as the linchpin for US efforts

in the Middle East later when the Shah dies

after being deposed or resigning.

Nixon is the only American president

to go to the Shah's funeral.

So he actually had a very friendly relationship with the Shah.

So what happens?

You get into the Jimmy Carter years.

Then you have a rising up of the Islamic radicals in Iran.

Where does the US Iran relationship

and with the Shah go from there?

Yeah, this is a real problem.

In fact, there's a great recent book called

King of Kings that talks about the Iranian revolution.

And how badly the US under Carter messed it up.

They just were not aware of all the

royaling troubles that were taking place

in that period.

In fact, Carter visits Tehran on New Year's Eve

right before the revolution.

And he calls Iran an island of stability

in one of the more troubled areas of the world.

Those words did not hold up well.

The Carter administration just wasn't aware

of all the problems that were happening.

There's this guy, Ayatollah Khomeini,

who's exiled to France.

And he has some enablers who are falsely telling the world

that he's going to be friendly to the West.

He's not going to be a problem.

All not true.

The Carter administration, the CIA,

had tapes of this guy, Khomeini,

and all the radical things he was saying,

they were in Farsi,

and the CIA put the tapes in a drawer,

and they never opened the drawer

or translated the tapes.

They didn't know how bad Khomeini was,

even though they could have,

had they just opened that drawer and looked at the tapes?

Did they welcome Khomeini?

Did they think he was going to be good somehow?

The Carter administration,

I guess, thought he was going to be okay.

Obviously not as close a relationship with the shop,

but they thought they could work with him.

And they were just wrong,

and obviously just a few months into this Khomeini takeover,

they take over the U.S. Embassy,

and the U.S. diplomats are held there for 444 days,

and that's really the thing more than anything else

that destroyed the Carter presidency.

Right, because he lost, of course, in 1980.

And he did attempt to rescue the hostages that didn't work out,

and then as we get into 1981,

after the election, Carter still tried to get the Americans freed,

but it didn't work out Reagan on his first day.

That's when the hostages were freed, correct?

Yeah, incredible frustration to Carter.

He spends all night,

and last night was presidency trying to get those hostages out,

and Khomeini is just a stick in the eye to Carter.

Make sure that the hostages are not freed until Reagan takes over.

There's two ways of looking at this.

Some people say that it was just a dig at Carter.

Others say that people were afraid,

or at least the Iranians were afraid what Reagan might do

if he were president, and they still were in captivity.

So either way, when Reagan takes over,

the hostages are released,

but that does not necessarily improve US relations with Iran.

When do the terror tentacles spread for the Iranians?

I mean, getting into, you know, Hezbollah,

and then, of course, Hamas and Gaza and all that.

How does that happen?

Yeah, it's really important to highlight this,

that Iran is the sponsor, the supporter, the adviser,

of organizations that hate America, try to destroy America,

like Hezbollah and Hamas,

and this really starts in 1983 in the Beirut bombing

when Hezbollah, which is again an Iranian-supported organization,

murders 241 US service members in a terrible truck bombing.

But that wasn't the only thing that they were doing.

They hijacked a commercial plane where they murdered a US Navy diver.

They took multiple Americans hostage in Iran,

and that was actually what set in motion the Iran-Kacher affair,

which was the biggest scandal of Ronald Reagan's presidency.

So from the minute they get into power,

they have this idea that they want to destabilize things

using proxies.

And the advantage of proxies, like Hezbollah and later Hamas,

is that Iranians don't suffer when the Hezbollah people are killed

or when the Hamas terrorists are killed.

The Iranians have this level of insulation.

And right now, what's going on

with the attack against Iran is after 47 years of this strategy

that's attacking the US through proxies.

Finally, the US is saying, you know what?

We're going to hit you back directly,

not just through the proxies.

So as we get through the first President Bush years

into the Clinton years, we get into the second President Bush.

He labels Iran part of the axis of evil.

What did that do?

Well, first of all, it called what it was.

I mean, Iran was a very problematic nation.

Still is, hopefully, where things are going to change more positively.

He tried to get the Iranians to the nuclear negotiating table

in his second term, but the real problem was that the Iranians

were fomenting trouble against US troops in Iraq.

600 Americans died because of roadside bombs

that the Iranians had provided for Iraqi terror networks

through Kasim Salamani, who Trump took out the Iranian al-Quds Force

was a constant irritant to the American efforts in Iraq.

But the Bush administration had other things on its mind.

It was bogged down by the words in Iraq and Afghanistan,

and it did little to address the Iranian problem

because they just didn't have the bandwidth for it.

President Obama did enter into a nuclear agreement

with Iran in his presidency.

It was undone by President Trump in his first term,

but Democrats have said that's a mistake.

I think the mistake was the Obama approach,

which was this kind of balance of power approach

where you kind of prop up the Iranians as one axis

against the Sunni states and Israel on the other side.

The truth is, if you have allies, you should support your allies,

and that is not what Obama was doing.

He was trying to work with the Iranians,

and he called Iranian President Hassan Rouhani,

which was the first contact between an American president

and any Iranian head of state.

It didn't do much good.

They had this so-called nuclear agreement

that didn't really stop them from developing nuclear weapons,

certainly didn't stop their missile program

and did not stop their use of proxies.

So it was really an ineffectual deal that President Trump pulled out of.

But the problem is you pull out of the deal

and you still don't have a way to stop them

from developing nuclear weapons.

So it was a bad deal, but we still needed to do something else

to stop them from getting nuclear weapons

and attaching them to ballistic missiles.

We already see that their ballistic missiles

can wreak havoc across the entire region.

Tevi Troy is a presidential historian,

the author of five books on the presidency,

former senior White House aide in the George W. Bush presidency,

senior fellow with the Ronald Reagan Institute.

Great to have you on the show.

Thanks so much for being here.

Thanks for having me.

Instagram teen accounts default teens

into automatic protections for who can contact them

and the content they can see.

Explore teen accounts,

and all of our ongoing work to protect teens online

at Instagram.com slash teen accounts.

Tomorrow on The Fox News rundown,

we speak with the Republican chairman

of the House Homeland Security Committee, Andrew Garberino.

As Democrats keep blocking funding for Homeland Security,

we talk about the ouster of Christine Ohm

as Homeland Security Secretary Ed,

of the conflict in Iran is raising concern

about the threat of domestic terrorism.

I'm Dave Anthony,

and this is The Fox News rundown from Washington.

news.com.