4 ms·

Army general says he's using AI to improve "decision-making"

FrustratedMonky 1y ago
The military using AI for decisions was one of the inflection points in AI 2027 https://www.youtube.com/watch?v=k_onqn68GHY https://www.youtube.com/watch?v=k_onqn68GHY https://ai-2027.com/ https://ai-2027.com/
- alganet 1y ago
  The movie Terminator 3: Rise of The Machines also predicted military involvement just before an AI apocalypse, in 2003. Why do people believe in the AI-2027 timeline and not Terminator 3: Rise of The Machines? That's because AI-2027 is not labeled as fiction, the categorization it actually deserves. Fiction can sometimes have predictive power, but that's not the case for AI-2027. Their goal does not seem to be creating art to entertain people.
  - FrustratedMonky 1y ago
    "predicted military involvement " Everybody and their brother predicts Military Involvement. What 2027 describes is subtle creeping into daily ordinary decision making. AI controlling drones is old news. This is more, at all levels, they can't decide to get a coffee without AI.
    - alganet 1y ago
      I don't know, dude. Maybe the general is cracking an internal joke that we, the public audience, can't quite grasp. For example, it could be that some company offered this kind of AI to the army, with those exact words, and they refused (which sounds incredibly reasonable). The interview would serve a purpose of deliberately mocking such offer, in public. There is almost no downside to this, as you already mentioned "everybody and their brother predicts Military Involvement [with AI]", and I quote you to illustrate that the idea that this is engrained in the popular culture (maybe by design). Companies such as Raytheon and other traditional military contractors have been researching AI (not necessarily chats) way before Sillicon Valley. They can (and probably do) offer consultancy to the army regarding which technologies are safe to employ for critical decision-making. I'm not defending them, by the way, just exploring some ideas around what possibly is going on here. Letting my imagination go a little beyond what the typical scenarios imply. What I know for sure, is that AI-2027 is not a serious prediction. It's a piece of fiction. Kind of in the vein of The Blair Witch, with some tones of realism to impress the public, and a little bit more sinister in the sense that it resists the idea of being interpreted as fiction (which I think it's a downer, they could have reached more acclaim by being artists than scammers).
      - FrustratedMonky 1y ago
        For weapons. There is already AI piloted F16's that win against human pilots. This is already here, now years old. For Analysis. Who knows what the military has access to. I was able to ask plain old GPT about India/Pakistan, and it was able to weed through media reports and come up with something pretty accurate. Maybe we wont launch nukes based on an LLM feedback. But it will definitely be used to summarize and analyze data into reports. Not sure why this seems like it wont be done. Military has many rungs of management that needs to push out reports just like a corporation.
        alganet 1y ago
        I'm sure it will be done but I don't buy the version neither the timeline that AI-2027 promotes. Compreende?
        FrustratedMonky 1y ago
        Yes. All predictions are fiction. I understand that. Arguing over 5 years or 30, and some future improvement that is not exactly the LLMs of today. Not sure we are disagreeing on the eventuality.
        alganet 1y ago
        Do not be mistaken, we totally disagree. Not all predictions are fiction. I would not try to fit AI-2027 into a news article, you did. See? Disagreeing heavily here.
        FrustratedMonky 1y ago
        "I'm sure it will be done " - alganet I must have misinterpreted this. I really don't want to argue around defending 2027. It is a speculative view. They say so themselves. But speculation on the future has been a valuable tool in the past. So not dismissing out of hand. These were researchers with some respectable credentials. Not simply fan fiction like Shades of Gray or something. As to news. There was an article, it matched a point in 2027, i made the connection. How is pointing out a similar connection or idea, invalid? 'fitting speculation into news' . Like no connections can ever be made?
ungreased0675 1y ago
I would bet my paycheck General Taylor is experiencing AI psychosis. AI is good for a lot of things, but military decision making isn’t one of them.
jleyank 1y ago
He could save power by just using dice and decision tables. There's a long history of civilian and military wargaming that's random but designed vs. real-world situations. Statistical token generation, however, doesn't strike me as a good bet. Hell, during one of the baseball games, Google AI led off their search results with "Seattle is not north of Toronto". It followed by having little real knowledge of historical baseball results. I can see it having problem due to training set with the latter, but damn, where countries and cities are located should be well-defined.
nis0s 1y ago
Sounds tricky if the LLM isn’t fine tuned to a dataset with examples of the kind of strategizing that you’d like to optimize for. Untrained LLMs are experts at average responses, and better strategies for anything require creativity, domain knowledge, and lateral thinking.

The military using AI for decisions was one of the inflection points in AI 2027 https://www.youtube.com/watch?v=k_onqn68GHY https://www.youtube.com/watch?v=k_onqn68GHY https://ai-2027.com/ https://ai-2027.com/

alganet 1y ago
The movie Terminator 3: Rise of The Machines also predicted military involvement just before an AI apocalypse, in 2003. Why do people believe in the AI-2027 timeline and not Terminator 3: Rise of The Machines? That's because AI-2027 is not labeled as fiction, the categorization it actually deserves. Fiction can sometimes have predictive power, but that's not the case for AI-2027. Their goal does not seem to be creating art to entertain people.
- FrustratedMonky 1y ago
  "predicted military involvement " Everybody and their brother predicts Military Involvement. What 2027 describes is subtle creeping into daily ordinary decision making. AI controlling drones is old news. This is more, at all levels, they can't decide to get a coffee without AI.
  - alganet 1y ago
    I don't know, dude. Maybe the general is cracking an internal joke that we, the public audience, can't quite grasp. For example, it could be that some company offered this kind of AI to the army, with those exact words, and they refused (which sounds incredibly reasonable). The interview would serve a purpose of deliberately mocking such offer, in public. There is almost no downside to this, as you already mentioned "everybody and their brother predicts Military Involvement [with AI]", and I quote you to illustrate that the idea that this is engrained in the popular culture (maybe by design). Companies such as Raytheon and other traditional military contractors have been researching AI (not necessarily chats) way before Sillicon Valley. They can (and probably do) offer consultancy to the army regarding which technologies are safe to employ for critical decision-making. I'm not defending them, by the way, just exploring some ideas around what possibly is going on here. Letting my imagination go a little beyond what the typical scenarios imply. What I know for sure, is that AI-2027 is not a serious prediction. It's a piece of fiction. Kind of in the vein of The Blair Witch, with some tones of realism to impress the public, and a little bit more sinister in the sense that it resists the idea of being interpreted as fiction (which I think it's a downer, they could have reached more acclaim by being artists than scammers).
    - FrustratedMonky 1y ago
      For weapons. There is already AI piloted F16's that win against human pilots. This is already here, now years old. For Analysis. Who knows what the military has access to. I was able to ask plain old GPT about India/Pakistan, and it was able to weed through media reports and come up with something pretty accurate. Maybe we wont launch nukes based on an LLM feedback. But it will definitely be used to summarize and analyze data into reports. Not sure why this seems like it wont be done. Military has many rungs of management that needs to push out reports just like a corporation.
      - alganet 1y ago
        I'm sure it will be done but I don't buy the version neither the timeline that AI-2027 promotes. Compreende?
        FrustratedMonky 1y ago
        Yes. All predictions are fiction. I understand that. Arguing over 5 years or 30, and some future improvement that is not exactly the LLMs of today. Not sure we are disagreeing on the eventuality.
        alganet 1y ago
        Do not be mistaken, we totally disagree. Not all predictions are fiction. I would not try to fit AI-2027 into a news article, you did. See? Disagreeing heavily here.
        FrustratedMonky 1y ago
        "I'm sure it will be done " - alganet I must have misinterpreted this. I really don't want to argue around defending 2027. It is a speculative view. They say so themselves. But speculation on the future has been a valuable tool in the past. So not dismissing out of hand. These were researchers with some respectable credentials. Not simply fan fiction like Shades of Gray or something. As to news. There was an article, it matched a point in 2027, i made the connection. How is pointing out a similar connection or idea, invalid? 'fitting speculation into news' . Like no connections can ever be made?

ungreased0675 1y ago

I would bet my paycheck General Taylor is experiencing AI psychosis. AI is good for a lot of things, but military decision making isn’t one of them.

jleyank 1y ago

He could save power by just using dice and decision tables. There's a long history of civilian and military wargaming that's random but designed vs. real-world situations. Statistical token generation, however, doesn't strike me as a good bet. Hell, during one of the baseball games, Google AI led off their search results with "Seattle is not north of Toronto". It followed by having little real knowledge of historical baseball results. I can see it having problem due to training set with the latter, but damn, where countries and cities are located should be well-defined.

nis0s 1y ago

Sounds tricky if the LLM isn’t fine tuned to a dataset with examples of the kind of strategizing that you’d like to optimize for. Untrained LLMs are experts at average responses, and better strategies for anything require creativity, domain knowledge, and lateral thinking.