Episode summary
Anthropic, the Pentagon, and the Future of Autonomous Weapons | Odd Lots
Odd Lots
· · Hosts: Tracy Alloway, Joe Weisenthal · Guests: Paul Scharre
Summary
The episode centers on the dispute between Anthropic and the Pentagon over how AI models can be used in war, especially around autonomous weapons and surveillance. The hosts argue that the immediate controversy is less about whether the military can use AI at all than about who gets to set the rules for lawful use, and whether companies can trust government assurances once the models are deployed.
Paul Scharre says an autonomous weapon is fundamentally a system that chooses its own targets, but he emphasizes that current military use is far short of that. He describes today’s battlefield AI as a mix of mature image-classification systems and newer large language models embedded in the Pentagon’s Maven workflow to help analysts sort imagery, signals, geolocation data, and logistics. In his telling, humans remain in the loop, but the real risk is that they become rubber stamps for large target sets.
Scharre also warns that the technology trendline is moving toward more capable, multimodal, and networked systems that could erode meaningful human control, especially in cyber conflict or fast-moving drone warfare. He argues that AI could also improve discrimination and reduce civilian harm if used as a warning or vetting layer, but he is skeptical that fully autonomous weapons are a near-term reality or that they can be safely governed without strong human oversight.
Key takeaways
- Paul Scharre says an autonomous weapon is a system that chooses its own targets, while today’s military AI mostly supports human planners rather than replacing them.
- The Pentagon is already using AI for image classification and for planning support through the Maven smart system, including with Anthropic tools in the Iran campaign.
- Scharre warns that humans can become nominally “in the loop” but still rubber-stamp AI outputs when target sets get large and fast-moving.
- The episode frames the Anthropic dispute as a fight over contract terms and policy boundaries, not a near-term push for fully autonomous killing machines.
- Scharre thinks AI could reduce civilian harm if used to flag risky strikes or suggest smaller munitions, but he says that still requires meaningful human judgment.
Notable quotes
“I think conceptually I think the distinction really is a weapon that is choosing its own targets okay on the battlefield”
“the large language model tools, whether it's Claude or other companies, can be valuable is there can be a way for a human to interact with that data”
“humans are pretty involved right now in terms of actually looking at the output from AI giving pretty specific guidance to the AI systems”
“What's at dispute here is a more fundamental disagreement about well, who sets the rules”
“I think that is a very serious risk. And so the like mental model that I have for this are things like flash crashes”
How the episode opens
Bloomberg Audio Studios podcasts, radio, news. — Hello and welcome to another episode of the OddLots podcast. I'm Joe Wisenthal — And I'm Tracy Aloway. So Tracy, we're recording this March 24th and of course almost all of our episodes lately have been about the war in Iran,
But what's interesting or what's a little weird is that just prior to the war, literally days or maybe hours, the biggest story in the world was actually about defense and you know the DoD. — That's right. So you are referring to anthropics and its disagreement to put it mildly with the department of war.
Yeah, exactly. This was the biggest story going right up on the eve of the war in Iran. And of course, obviously, there was this contract and anthropics technology was used by the defense department. So, it was not a disagreement about the use of AI per se in war.
But the question of the degree of to which AI could be used for autonomous weapon systems on their own without a human in the loop — And surveillance, that was — Surveillance and surveillance. This was another key element.
But you're right. So, we've heard this expression autonomous weapons pop up more and more, especially in recent days. And I have a lot of questions over what exactly that means because my impression is the US military certainly has been using AI for some time.
And so, we're really talking about degrees here of autonomy, right? And so yes, if you if you think about an autonomous weapon, I think your mind could go fully Terminator and you know there's like a murder robot out there that's making its own decisions on which people or places to target and then you get levels below that, right?
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- autonomous weapons
- ai in warfare
- pentagon ai policy
- military surveillance
- ethics of autonomous systems
Companies and tickers mentioned
- Anthropic
Watch the episode
Watch “Anthropic, the Pentagon, and the Future of Autonomous Weapons | Odd Lots” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Odd Lots
-
How a Chicago Restaurant Has Stayed Open For Nearly Three Decades | Odd Lots -
How Airlines Actually Hedge Higher Fuel Prices | Odd Lots -
Robert Friedland on the World's Monumental Shortage of Copper | Odd Lots -
Brad Setser on the US's Unusual Japanese Yen Intervention | Odd Lots -
The Tungsten Market Is Warning of an Upcoming War | Odd Lots -
These Are the Sharps Actually Making Money on Prediction Markets | Odd Lots
Odd Lots, this episode and its artwork belong to Bloomberg. Octagon is not affiliated with the show; this page is an independent summary.
