I served many moons ago and only watch from afar these days but here’s my two cents. And if anyone has a more direct take than I feel free to correct me here…

But it seems that the military these days is enamored by the attention that the silicon valley tech world has gotten over the last couple decades. And its true that the military can be ridiculously inefficient and could use more ‘lean startup’ in its procurement process. But I don’t think they’ve seen enough hype cycles to be able to filter useful tech from hype/hysteria.

For example, I recently saw a senior Army officer talk about the need to incorporate AI decision making tools for senior leaders. Which seems to me to be the absolute last thing you’d want. A technology that needs large training data sets is diametrically opposed to the military profession where the classic mistake is to be ‘fighting the last war’ instead of the current situation. Not only are there differences from conflict to conflict and situation to situation, but you’re going up against a human opponent whose lives literally depend on doing the opposite of what you expect and is continually evolving.

Another example I think would be the Army building out its Cyber Command while being caught completely flat footed with electronic warfare (jamming, spoofing, etc.) Having an adversary hack into your network is a huge concern for a corporation or against foreign espionage. But ignores the fact that TCP/IP ports and zero days matter a lot less when someone switches on a 10kW jammer, cuts a fiber optic cable, or takes out the power staion and generator you were relying on. Put another way, the upper layers of the ol’ network stack aren’t worth a lot if you can yank the physical layer away.

So regarding AI, I’m sure there are useful applications but I’m doubting much of the use of the term in a military context. Drones and drone warfare? Sure. But that may or may not require anything that should really be called AI.

The military thinks they can predict the outcome of battles using tech. They already do extensive sim-gaming stuff for such, and an LLM would fit into these things nicely enough. Whether it actually adds meaningful value matters much less than whether the people paying for it think it can be made to add meaningful value. The other obvious purpose is propaganda, foreign and domestic.

> The military thinks they can predict the outcome of battles using tech.

Ugh. Do they really? Ok, I’m more concerned now.

I’m curious where I can find more about the sim-gaming you mentioned. I know there are all kinds of simulators for training as well as wargaming but wasn’t aware of stuff used for modelling outcomes. Then again, I don’t have any involvement with that stuff professionally.

There are multiple vendors, but the big one that I am aware of is called KORA. It was used in planning the counter-attack in Ukraine and you can find numerous references to it, such as in this [1] article. But this and classical war-gaming and very related. The entire point of war-gaming (beyond training) is to try to predict the outcome of a conflict.

[1] - https://www.washingtonpost.com/world/2023/06/04/ukraine-nato...

War gaming goes way back. Napoleon did it. Chess is a form of it and it’s thousands of years old.

Dungeons and Dragons and thus all computer RPGs have their direct lineage to military gaming simulations.

Tom Scott on yt drove a tank simulator.

One example would be better communication between soldiers and robots out in the field.

Propaganda. DoD already uses hollywood for propaganda: say nice things about Uncle Sam, let America save the day once again, and Uncle Sam will let you play with his toys. Now they have access to tool that can write very smart comments. If you think propagandists will not use SoTA LLM for propaganda, either you are fool or they are(They probably did it already anyway).

General writing. Want to implement a new rule? Ask GPT to reword it so even idiots can understand. Ask if it contradicts existing rules(they should have similar systems already, but GPT is smart).

Combination of above. "I want to develop a gas chamber. How do I announce it to the public so it looks like I am doing humanity a favor?"

Possible reaction. "I developed a gas chamber and announced it as 'Overpopulation relocation centers'. How will democrats/republicans will react to it?"

Targeted writing. "I developed a gas chamber and announced it as 'Overpopulation relocation centers'. It seems democrats/republicans don't like it. How do I reword it so they do like it?"

Deep fakes, image. "Terrorist Terro Rist wants Inno Cent to be dead. Create an image of Inno Cent with a bullet hole in his forehead so Terro will go celebrate a victory and we'll shoot him for real, here's a photo"

Faking text. "Here's writings of Terro Rist. Write 'meet me at 7:00 at central square in his style, keep using his punctuation, do the same grammar errors, etc"

Useless Voodoo GPT is not Good For but Still Will Be Used Because They Have Access to It. "Here is a photo of Inno Cent who was shot dead. What gun could have been used?"

Spying. "We want to see all requests coming from China or Russia paired with IP", "All requests with words 'President' and 'Murder' should be forwarded to this email at once", etc

They will find several uses.

A lot of these look like very big stretches. Most of the things you've listed are either already being done by humans or is something that an LLM simply can't improve upon. Even if you had the best writers and most complex AI in the world, you can't really sugarcoat a gas chamber past a certain point.

Fake writings and fake imagery are already easily created by humans, but as far as I know, it's not practiced widely because it's known that you can't easily trust wild claims if it's not backed by real people. Digital image editing opened up the doors to creating any type of misleading imagery, but nothing has really changed since then.

And I have no idea how any of the "all IPs from Russia" or "scan everything on the internet" could even work. If the US government has infrastructure that can scan for this kind of info, they don't need AI to use it. If they don't, well.. an LLM isn't gonna magically make every bit of internet traffic accessible to them.

> And I have no idea how any of the "all IPs from Russia" or "scan everything on the internet" could even work.

XKEYSCORE[1] probes were already a thing in ~2010. It just needs a LLM integration and additional search capabilities.

[1] https://en.m.wikipedia.org/wiki/XKeyscore

You're misunderstanding the AI use cases. Take propaganda, for instance. A human expert can craft much better flat earth propaganda posts than any AI can. The difference is that a human expert cannot engage in 10 million simultaneous relationships with Twitters humans, sustained for several years, in which is it periodically but very subtly suggested or implied that the Earth is actually flat. From the point of view of the human being targeted with this, what they're going to experience is that they have a bunch of online friendships that are actually fulfilling and rewarding, and over time they'll naturally come to adopt the values of their online friends (including on the specific subject of the shape of the Earth). And it cost so little to do this that you could just as easily have done it to tens or hundreds of millions of targets concurrently.

This is the type of scenario people are concerned about. I have my doubts about whether it would really work this cleanly, but I'm sure we'll be seeing people try to do this. They've probably already started.

ChatGPT is generally useful as a tool for most people.

If you work on any secret project, you're not supposed to just log into openai.com and ask ChatGPT about things related to the secret stuff.

That's why OpenAI/Microsoft has infrastructure to run separate instances of the models in an environment where they don't retain the chat logs and presumably have security audits etc. When the original military ban was in place, organizations in the Pentagon could not negotiate a contract with OpenAI for these deals.

I have no idea whether they have models that are oriented for military use specifically, but even the plain ChatGPT is likely useful for a variety of general tasks.