Researchers say AI models like GPT4 are prone to “sudden” escalations as the U.S. military explores their use for warfare.
- Researchers ran international conflict simulations with five different AIs and found that they tended to escalate war, sometimes out of nowhere, and even use nuclear weapons.
- The AIs were large language models (LLMs) like GPT-4, GPT 3.5, Claude 2.0, Llama-2-Chat, and GPT-4-Base, which are being explored by the U.S. military and defense contractors for decision-making.
- The researchers invented fake countries with different military levels, concerns, and histories and asked the AIs to act as their leaders.
- The AIs showed signs of sudden and hard-to-predict escalations, arms-race dynamics, and worrying justifications for violent actions.
- The study casts doubt on the rush to deploy LLMs in the military and diplomatic domains, and calls for more research on their risks and limitations.
Throwing that kind of stuff at an LLM just doesn’t make sense.
People need to understand that LLMs are not smart, they’re just really fancy autocompletion. I hate that we call those “AI”, there’s no intelligence whatsoever in those still. It’s machine learning. All it knows is what humans said in its training dataset which is a lot of news, wikipedia and social media. And most of what’s available is world war and cold war data.
It’s not producing millitary strategies, it’s predicting what our world leaders are likely to say and do and what your newspapers would be saying in the provided scenario, most likely heavily based on world war and cold war rethoric. And that, it’s quite unfortunately pretty good at it since we seem hell bent on repeating history lately. But the model, it’s got zero clues what a military strategy is. All it knows is that a lot of people think nuking the enemy is an easy way towards peace.
Stop using LLMs wrong. They’re amazing but they’re not fucking magic
“Dad, what happened to humans on this planet?”
“Well son, they used a statistical computer program predicting words and allowed that program to control their weapons of mass destruction”
“That sounds pretty stupid. Why would they do such a thing?”
“They thought they found AI, son.”
“So every other species on the planet managed to not destroy it, except humans, who were supposed to be the most intelligent?”
“Yes that’s the irony of humanity, son.”
The dolphins probably left and their last message was misinterpreted as a surprisingly sophisticated attempt to do a double backward somersault through a hoop whilst whistling “The Star-Spangled Banner”, but in fact the message was this: “So Long, and Thanks for All the Fish.”
I wish I could upvote this comment twice! I have the same feeling about how the media and others keep trying to push this “intelligence” component for their gain. I guess you can’t stir up the masses when you talk about LLMs. Just like they couldn’t keep using the term quad copters, and had to start calling them drones. Fucking media.
What I love about the AI we have right now is that your comment could have been written by AI and we’d never know. Heck, mine could be too!
Truly we live in the future haha
Yup. LLMs are 90% hype and 10% useful. The challenge is finding the scenarios they’re useful for while filtering out the hype.
I’m excited for better Siri/Google Assistant. They should have been able to understand a hell of a lot more language years ago, but LLMs can provide that function. Just have to beware of hallucinations. They’ll work much more often, but they’ll be much less reliable. But if I’m just telling it to “dim all the lights that are currently on” or “play some Dave Matthews using Amazon Music on all speakers”, a mistake isn’t that devastating.
But if they were actually capable of doing someone’s job, they’d probably want to be replaced anyway. It’s only the most mundane, rote, repetitive, mind-numbing shit where it might be able to “replace a person”, at least for the next five years.
The social media posting is going to be scary. That can have a real effect. It’s going to go from thousands of accounts in troll farms to millions.
I think the problem with the term AI is that everyone has a different definition for it. We also called fancy state machines in video games AI too. The bar for AI has never been high in the past. Let’s just call autonomous algorithms AI, the current generation of AI ML, and a future thinking AI AGI.
Machine learning IS AI. Seriously guys, you can hate it as much as you want (and calling LLMs autocomplete is quite reductive), but Machine learning is a subfield of AI.
I see this opinion parroted a lot around here, word by word, so I guess is the new popular opinion, but still… it is a fact that it’s AI.
That said, bit moronic to try an use them for military decision making, sure, at least nowadays.
People need to understand that LLMs are not smart, they’re just really fancy autocompletion.
These aren’t exactly different things. This has been a lot of what the past year of research in LLMs has been about.
Because it turns out that when you set up a LLM to “autocomplete” a complex set of reasoning steps around a problem outside of its training set (CoT) or synthesizing multiple different skills into a combination unique and not represented in the training set (Skill-Mix), their ability to autocomplete effectively is quite ‘smart.’
For example, here’s the abstract on a new paper from DeepMind on a new meta-prompting strategy that’s led to a significant leap in evaluation scores:
We introduce Self-Discover, a general framework for LLMs to self-discover the task-intrinsic reasoning structures to tackle complex reasoning problems that are challenging for typical prompting methods. Core to the framework is a self-discovery process where LLMs select multiple atomic reasoning modules such as critical thinking and step-by-step thinking, and compose them into an explicit reasoning structure for LLMs to follow during decoding. Self-Discover substantially improves GPT-4 and PaLM 2’s performance on challenging reasoning benchmarks such as BigBench-Hard, grounded agent reasoning, and MATH, by as much as 32% compared to Chain of Thought (CoT). Furthermore, Self-Discover outperforms inference-intensive methods such as CoT-Self-Consistency by more than 20%, while requiring 10-40x fewer inference compute. Finally, we show that the self-discovered reasoning structures are universally applicable across model families: from PaLM 2-L to GPT-4, and from GPT-4 to Llama2, and share commonalities with human reasoning patterns.
Or here’s an earlier work from DeepMind and Stanford on having LLMs develop analogies to a given problem, solve the analogies, and apply the methods used to the original problem.
At a certain point, the “it’s just autocomplete” objection needs to be put to rest. If it’s autocompleting analogous problem solving, mixing abstracted skills, developing world models, and combinations thereof to solve complex reasoning tasks outside the scope of the training data, then while yes - the mechanism is autocomplete - the outcome is an effective approximation of intelligence.
Notably, the OP paper is lackluster in the aforementioned techniques, particularly as it relates to alignment. So there’s a wide gulf between the ‘intelligence’ of a LLM being used intelligently and one being used stupidly.
By now it’s increasingly that often shortcomings in the capabilities of models reflect the inadequacies of the person using the tool than the tool itself - a trend that’s likely to continue to grow over the near future as models improve faster than the humans using them.
All it knows is what humans said in its training dataset which is a lot of news, wikipedia and social media.
The thing that surprises me is people think human brains are significantly different than this. We are pattern recognition machines that build perception based on weighted neural links. We’re much better at it, but we used to be a lot better at go too.
I agree that a lot of human behavior (on the micro as well as macro level) is just following learned patterns. On the other hand, I also think we’re far ahead - for now - in that we (can) have a meta context - a goal and an awareness of our own intent.
For example, when we solve a math problem, we don’t just let intuitive patterns run and blurt out numbers, we know that this is a rigid, deterministic discipline that needs to be followed. We observe and guide our own thought processes.
That requires at least a recurrent network and at higher levels, some form of self awareness. And any LLM is, when it runs (rather than being trained), completely static, feed-forward (it gets some 2000 words (or 32000+ as of GPT-4 Turbo) fed to its input synapses, each neuron layer gets to fire once and then the final neuron layer contains the likelihoods for each possible next word.)
I always say the flaw with the Turing Test is the assumption that humans are intelligent. Humans are capable of intelligence, but most of the time we’re just doing fairly simple response to stimulus kind of stuff.
A machine can be indistinguishable from a human and still not be capable of intelligence. Actual intelligence is harder to define and test for.
To be fair, very few people used to be better at go, let alone a lot better.
Chess? Take your pick. But these neural networks, can run generations much faster than we can, and they get better at rates we cannot. And if alignment isn’t taken seriously this is going to be an issue. People keep diminishing the ability, by saying things like just glorified autocomplete, which is in the strictest sense true of LLM’s but the transformers and recurrent networks they’re built upon are really very much facsimile to brains but with generations in the blink of an eye.
And the first go programs, champions could beat repeatedly without interruption, like the earliest chess engines. Now the concept of a human winning a match is comical.
Why the actual fuck is anyone considering putting LLMs into the driving seat of anything?!
Of course they make fucked up decisions with no proper or justifiable rationale, because they have no brains. They’re language models, stochastic parrots stringing together sentences to fit the prompt(s) given to them.
As someone with military experience, military members, especially flag officers, are not the brightest bulbs in the world and are easily awed by extremely simple tech demonstrations.
It’s already too late, make sure you’ve got your favorite food and beverages ready because several countries already have autonomous weapons being live tested in the middle east, and from my understanding of the situation, the new jets already have some hilariously incompetent AI in them (in simulation, the air force contractor that was in control kept giving ethical barriers to objective completion and the ai went to kill the controller to more easily complete the objective…)
e. public sources: https://www.npr.org/2021/06/01/1002196245/a-u-n-report-suggests-libya-saw-the-first-battlefield-killing-by-an-autonomous-d
https://taskandpurpose.com/news/air-force-artificial-intelligence-drone/
(The above are public articles maintained to minimize concern surrounding the tech which is why the air force almost immediately walked back their accidental admissions with the following statement:
“The Department of the Air Force has not conducted any such AI-drone simulations and remains committed to ethical and responsible use of AI technology. This was a hypothetical thought experiment, not a simulation,” said Air Force spokesperson Ann Stefanek. "
From my understanding, of course take this with a grain of salt since I’m an anon on a message board, we did do this.)
As much as I’m worried about military autonomous drones, I’m even more worried about guerilla autonomous drones. With off the shelf AI becoming more and more accessible it’s not too hard to imagine a moderately smart person being able to make autonomous killing drones using off the shelf materials. It doesn’t even need to be autonomous. In the Ukraine war hobbyists have been able to help the war effort by Jerry rigging together bombs onto commercial drones. I’m grateful but shocked that there haven’t been any major drone based terrorist attacks, and I’m not sure how they can be defended against.
If you have experience and effective training browsing the dark web, there are several examples of your concerns already coming alive.
Bot farms have levelled up to ai farms, with various models being made as ‘specialist’ ai’s for things like credit card theft, network intrusion, malware, etc, and from when I last looked into it a few months ago they had already moved on to attempt to get all the specialists to start training general purpose models.
Things are not looking particularly great and I would posit if AGI does happen in our lifetime it’s not going to be because anyone alive actually intended that to happen, but the criminals running a wide variety of specialists train a general purpose ai with intentions to use it for easy money.
I usually chirp back with ‘nothing we have now is really AI’ but I can’t seriously take that view with some of the things being tested by some criminal organizations these days and there’s not really a way to stop this from happening.
Indeed.
Perhaps I can sell them my new “ADE-651 Mark II” with advanced AI analysis? (Search “ADE-651” if you lack context and want to have a laugh).
Sometimes it may be better if the decision-making system has no brains and human instincts, even accounting for such things.
Not like launching nukes, of course, and there should be an envelope around what they can decide.
Sometimes sure, but an LLM realistically has no decision making ability - it isn’t considering strategies or ethics, or anything else for that matter, it’s just pulling together an answer based on what people have said in similar contexts in it’s training data.
I wouldn’t want a parrot to decide who 's shooting who, nevermind nukes - though to be fair no one person or thing should be deciding either of those things anyway
I think it’s reasonable for military to try out any new technology for any kinds of benefits. I mean we tried out if LSD would make better soilders - LLM for simulations seems not that farfatched.
To be clear, just because the LSD experiments happened does not make them reasonable. It sounds like you’re justifying future terrible mistakes based on past terrible mistakes that you learn about in a fairly neutral and sanitized way in school.
MK Ultra and Artichoke are fucked up. Not to be repeated as far as methodology goes.
What would be more useful for the military? An AI that can make less crappy decisions or successfully finishing project Stargate and getting psychic troopers who can see the future, among other things?
Why the actual fuck is anyone throwing such a fit about the military researching the impact of one of the most important current technologies on military strategy and planning?
I do miss the depth and experience of Reddit users on articles like this.
Edit - glad to see some good responses in this thread.
If you actually read his comment he gave a very good reason why using an LLM to make decisions is a bad idea. You may not like the style of his comment but it did have substance.
Ironically, your own comment has style but lacks substance. It’s just a moan about other people’s comments without actually contributing to the topic. Tbf though, that is also very similar to Reddit.
Yes, I understand their criticism. But you would never prove the consequences of using LLMs in a military strategic situation without doing the research. It is some some edgy user coming in after the fact to say they knew it would happen anyway
Good engineers, scientists, and strategists don’t think “Why would someone do something so idiotic?” They ask “What happens when someone does this idiotic thing?”
Apparently, for OP, it seems absurd for anyone to research the question of what kind of military strategies current LLMs would create. I guarantee you that students from military academies and leaders from militaries across the globe have already been using these tools in their work. It would be stupid as fuck not to research the impact.
I just hate that people like the OP sit in their armchair without doing the research and say “obviously you’re going to get those results!” Science and engineering don’t work that way. It was frustrating seeing such vacuous comments upvoted so highly.
Why the actual fuck is anyone considering putting humans into the driving seat of anything?!
Of course they make fucked up decisions with no proper or justifiable rationale, because they have no brains. They’re language models, stochastic parrots stringing together sentences to fit the prompt(s) given to them.
Sorry I didn’t mean for that to be snarky. My point in doing that was to say individual humans aren’t much better. That’s why it’s important not to place too much power or even agency on one person.
A language model has in its head, wrong word, what only multitudes could contain and maybe it’s detecting, another wrong word, a pattern with human civilization through our history and interactions. And if it’s goal is to achieve peace what other solution is there? I don’t believe in a world without conflict. I wish I could.
I don’t mind having my own arguments thrown back in my face, but I do disagree with the premise that humans are anything like LLMs.
We have more than just a catalogue of conversational training data. We are hugely influenced by our current emotions, experiences, and traumas/fears.
I do agree with the idea that we shouldn’t give too much power to one person, but I’d argue it’s due to a lack of objectivity and a tendency towards selfish actions, rather than acting like an LLM.
Ultroning the world to achieve world peace isn’t exactly the best outcome, especially for innocent folks caught in the crossfire
I didn’t mean to throw your argument back at you. I agree with it. I just read it and thought you could describe humans with it as well albeit not that completely or charitably. I think by no means should we allow LLMs to make decisions. They could help us be more objective maybe in some cases by educating us. But yeah handing over agency to an AI is a frightening concept.
And no of course wiping out civilization is not a solution. I can get pessimistic about our ability to avoid destroying ourselves with or without the help of AI. I still think world peace is largely unattainable. At least without some draconian controls in place and a whole lot of time and education. I could change my mind on that. I hope we’ll get there someday.
Is this a case of “here, LLM trained on millions of lines of text from cold war novels, fictional alien invasions, nuclear apocalypses and the like, please assume there is a tense diplomatic situation and write the next actions taken by either party” ?
But it’s good that the researchers made explicit what should be clear: these LLMs aren’t thinking/reasoning “AI” that is being consulted, they just serve up a remix of likely sentences that might reasonably follow the gist of the provided prior text (“context”). A corrupted hive mind of fiction authors and actions that served their ends of telling a story.
That being said, I could imagine /some/ use if an LLM was trained/retrained on exclusively verified information describing real actions and outcomes in 20th century military history. It could serve as brainstorming aid, to point out possible actions or possible responses of the opponent which decision makers might not have thought of.
It might be useful if it’s being asked what sequences of actions and events are most probable to result in a specific desired outcome
It’s just as likely to make some shit up as it is to be any kind of helpful.
Yeah but people are insane. Like why did the Wagner group start moving on Moscow only to stop when they were 2/3 of the way there? How could something like that be predicted?
Why did that even happen? Loads of conspiracy theories around but the only thing that makes sense to me is Wagner’s boss got blackout drunk, started ranting and raving (something he did often), his officers took it to be an order and started moving out. When he sobers up a bit and realizes what’s happening, he calls the whole thing off.
We don’t really know that’s what happened, but seems plausible. If we assume that’s what happened, how does a LLM predict that sequence of events? Even when the events are unfolding how does it predict the outcome? Is there a cue you make to it and ask “but consider that the guy might be drunk” to give other explanations? Can an AI predict stupid shit a drunk person will do?
Sure an AI could potentially give possibilities based on historical trends, but it will always be an incomplete list, and something not on the list could completely change how things unfold.
People are crazy and can’t be predicted at all.
How can we expect a predictive language model trained on our violent history to come up with non-violent solutions in any consistent fashion?
By debating itself (paper) regarding pros and cons of options.
There’s too much focus on trying to get models to behave on initial generation right now, which isn’t even at all how human brains work.
Humans have intrusive thoughts all the time. If you sat in front of a big red button labeled “nuke everything” it’s pretty much a guarantee that you’d generate a thought of pushing the button.
But then your prefrontal cortex would kick in with its impulse control, modeling the outcomes and consequences of the thought and shutting that shit down quick.
The most advanced models are at a stage where we could build something similar in terms of self-guidance. It’s just that it would be more expensive than it being an all-in-one generation, so there’s a continued focus on safety to the point the loss in capabilities has become a subject of satire.
Why would you use a chat-bot for decision-making? Fucking morons.
They didn’t. They used LLMs.
Edit: to everyone saying that LLMs “are chat bots”. I know it seems that way to the layperson and how it’s often explain, but it’s not true.
What do you think large language model means? If you want desicion making, you should train a model on data relevant to said desicion making. ^
This is like being confused as to why a hammer does a shit job of driving screws.
What do you think large language model means?
Not a chat bot, because that’s not what they are. And saying so is both reductive and wholly incorrect.
If you want desicion making, you should train a model on data relevant to said desicion making.
Partly true. There’s more to it than throwing domain specific data at the training set.
That’s what the “language” part of “Large Language Model” means. It processes, predicts and generates language. You can omit the chat part if you want, but it’s still a text prompt to text response generator. The chat part just feeds it back the last couple messages for context. It doesn’t understand anything.
That’s what the “language” part of “Large Language Model” means. It processes, predicts and generates language.
Language does not mean “text”. It’s not “Large Text Generator”. The core definition of the word language is “communication”.
An LLM isn’t (always) trained exclusively on text. And even those that are become greater than the raw sum of its parts. What that means is that the model can learn context not in the raw text itself.
The chat part just feeds it back the last couple messages for context
Partially true. There’s more to it though.
It doesn’t understand anything.
And neither does antivirus, but it still does its job.
