Classified estimates show the NSA is paying billions to test AI models

bhouston

"Testing"? Sure, Jan. They would have to be incompetent if they were not using them both for security penetration as well as social engineering / election meddling as everyone else is.

hosel

Why wouldn’t they be? How is this surprising. A drop in the bucket.

paimapi

I feel like any AI regulation passed wouldn't have enough teeth to rein in the NSA because not even the Fourth Amendment seems applicable to the agency. let's not forget the ex-post-facto legalization of mass wiretapping and surveillance [0] with both establishment political parties supporting it on 'national security' concerns [1]. whatever monitoring or testing they're doing today will be similarly hand-waved away in the name of security theater this is just how the empire has always operated - we have always been at war with Eastasia [0] https://en.wikipedia.org/wiki/NSA_warrantless_surveillance_(... [1] https://www.eff.org/deeplinks/2012/10/fact-check-obamas-misl...

rdtsc

I guess I would be more surprised if the headline said "turns out NSA dropped the ball on AI and haven't even fired up a single LLM instance yet". It's mostly about "testing" and "billions". But I am sure "testing" here is in quotes just like I wrote it. They are not going to get an itemized list unless it's coming from Snowden 2.0 - We "tested AI" on this Russian nuclear silo command and control center. It worked out pretty ok.

rand846633

Does anyone here happen to know the ballpark number of how much the three letters agencies are currently paying OpenAI/antropic/who ever per year/month..?

clueless

My guess is, the US government (aka DJT) won't really acknowledge AI risks until folks at NSA start to freak out while using latest models for various tasks...

conorcleary

Yeah at some point they're going to seriously undervalue their own currency.

autoexec

Obviously. The amount of data they hold on the American public alone is massive. They can't possibly read it all, they need a system that can pull out "dangerous" people, and find dirt on whoever they want to target, and AI will happily do that, badly. I just hope they actually take the time to review what it gives them for accuracy so that they don't end up doing the NSA equivalent of bombing and murdering a bunch of elementary school children. https://en.wikipedia.org/wiki/2026_Minab_school_attack The NSA aren't the only ones with massive troves of very personal data on large numbers of people. Many companies are excited to use AI to help them use all of that information to their maximum advantage and you can bet that they won't care about spending the time/money on perfect accuracy as long as over the long run they get positive results.

binlog

Willing to bet any amount of money that these testing contracts are going to companies that the Trump family has invested in.

ck2

the day NSA cracks AES the world is going to get soooo weird I wonder how long it takes to decrypt 20 years worth of stored internet traffic day per year? week? certainly not a month per year everyone is focused on "AI" but there have been some significant quantum computer advancements

jawiggins

An underappreciated aspect of LLMs is how good they are at zero-shot classification. I'd expect this is extremely good at monitoring communications streams, particularly in foreign languages.

OutOfHere

This should remind us why it is inappropriate to pay a dollar more in taxes than you absolutely minimally have to by any means necessary. A fraction from every dollar you pay goes toward taking away your rights. You pay to have yourself be oppressed. I will note here two words that the NSA really likes to monitor across all devices: Trump, Monero.

9cb14c1ec0

Testing, yea right. No doubt they have gone on an epic AI assisted hacking spree. Gotten get persistence in all the things before vulnerabilities start getting patched.

vjvjvjvjghv

What do they test? In general, I don't understand the whole "AI safety" stuff. What are the scenarios we need to be protected from and how is this going to be accomplished?

passive

I don't think it's properly understood how directly Trump's administration is using taxpayer money to inflate the AI bubble. This is a small part of that. As a means to grift directly, but also for leverage over the tech industry, and possibly the entire economy. People hate data centers, but if they think AI regulation will crash the economy, will they vote Trump?

andreapaiola

Ok, I rewatch Person of Interest

JoeOfTexas

Meanwhile, I'm over here hesitating to dump $10 into yet another LLM model to test their API.

catchnear4321

Is this a good time to remind everyone about that data center in Utah that got everyone riled up and then went into the memory hole?

jmyeet

Companies inevitably reach a state where the only way to satiate the need for ever-increasing profits is by: 1. Raising prices; or 2. Cutting costs. Every use of AI falls into one of these. Demand pricing? Raising prices. Replacing people with AI systems? Cutting costs. But I really have to add a third to this list: 3. Weapons. Recently there's been reporting on how the Minot school bombing happened at the onset of the Iran War eg [1]. Humans normally have to be involved in target selection and the so-called "kill chain" as a safeguard. There are open questions here about how this happened. It's unclear if the human verification failed here or was bypassed by new policy. As another example, there are the AI targeting systems (eg Lavender, Where's Daddy?) used in Gaza [2]. So if you're working for any of these companies (eg OpenAI, Anthropic, SpaceX), you need to realize what it'll be used for: to more efficiently kill people. It's no different to working for RTX, Boeing or Lockheed Martin. As for the NSA, I've always been skeptical about the claims that the NSA records everything, simply because of the infrastructure required to do that. Networking, storage, physical server space, water for cooing, etc. Also, what good is it because nobody can possibly review it? But AI to the rescue! It's entirely reasonable to assume that every communication at some near future point will be intercepted, processed, tagged and logged by some government AI system. [1]: https://www.theguardian.com/news/2026/mar/26/ai-got-the-blam... [2]: https://www.972mag.com/lavender-ai-israeli-army-gaza/

hettygreen

So is this what all the datacenters are for?