Thread

Pentagon says overreliance on AI contributed to missile strike on Iran school

974 points · 553 comments · devonnull

legitster · 1790105097 · focus · HN ↗
> It found the U.S. “failed in its obligation to do everything feasible to verify” that the school was a military objective and that the failure “went beyond mere negligence.” The report said the United States “directed the strikes at the building of the school while being aware of a substantial risk of striking a civilian object and acting recklessly as regards the possibility that this would happen.”
Reading the details, "AI" doesn't really seem like the culprit -it's a scapegoat.
The intelligence that it was no longer a military target never entered the target database, the team that was responsible for vetting the target list was gutted, and said team was never even consulted.
The White House wanted 1000 targets and pulled from their database without any due diligence. Whether it was an AI call or an SQL query - this was from pure human maliciousness and incompetence.
1. beloch · 1790109043 · focus · HN ↗
  "In the final days of February, the Trump administration demanded an overwhelming aerial assault. More than 1,000 Iranian targets were hit within the first 24 hours alone. That order compressed the time frame for finalizing targets, the officials said, instilling an urgency that — combined with gaps in underlying information about the site, cuts in civilian-protection personnel and an overreliance on artificial-intelligence technology — set the stage for the deadly error."
  -------------------
  Maven is an untested tool for rapidly selecting targets under time pressure during a war. It's reason for existing is speed. If you're in a shooting war, the ability to select good military targets faster than your enemy, but not necessarily with zero error, is an advantage.
  Planning a first strike is a different situation. Time pressure should not be a significant factor. Humans should have been doing their due diligence on every single target. If an untested tool for rapid mid-battle target selection was relied upon completely because of a hasty executive order, then the responsibility falls entirely on the executive.
  1. cgio · 1790113999 · focus · HN ↗
    I think that’s the important distinction. Mistakes under pressure do happen. Unnecessary pressure is not a mistake, though, it’s a crime in this context. They could start planning earlier or commit to action later.
    ygouzerh · 1790150387 · focus · HN ↗
    I feel as well that it's the strongest reason. Even without AI, time pressure statistically already increase the error rate
  2. pbhjpbhj · 1790116305 · focus · HN ↗
    Well the order wasn't lawful, Congress hadn't countenanced a war.
    Unfortunately the military all turned their backs on their oath to uphold the constitution (Again! Deploying against USA cities and kidnapping foreign heads-of-state at the whim of Trump is also not lawful). At that point it is hard to see how your democracy can be resurrected.
    iugtmkbdfil834 · 1790118993 · focus · HN ↗
    << At that point it is hard to see how your democracy can be resurrected.
    Oddly, here I am more optimistic. The antipathy towards Tman rivals Musk and few other public personas and that, previously somewhat small crowd, appears to be growing. Historically speaking, I would say it is unlikely based on what happened ( or did not happen ) with Bush, but may actually unite enough people to punish its executive power center. I will admit that going into this year, this was not even close to my anything close to my bingo card. Then again, neither was Greenland, import war with Canada, Iran war and now open talk about banning diesel exports.
    kadoban · 1790138779 · focus · HN ↗
    > The antipathy towards Tman rivals Musk and few other public personas and that, previously somewhat small crowd, appears to be growing.
    That would make me optimistic as well if dictatorships governed with the consent of the populace. They do not.
    iugtmkbdfil834 · 1790150855 · focus · HN ↗
    I don't want to get philosophical on you, but what is consent, exactly. For example, in US, most would claim the current administration has consent of the populace to govern and yet, somehow, priorities of said populations are nowhere close to being met suggesting that either the population's consent is either imaginary, temporary, conditional, a phrase or an more interesting mix of those. And amusingly, we do not even have to search hard for it, because this administration has a funny habit of saying quiet part out lout ( <a href="https://www.globalsecurity.org/military/library/news/2025/12/mil-251219-state01.htm" rel="nofollow">https://www.globalsecurity.org/military/library/news/2025/12... - "We don’t do foreign policy by poll, and our job is to carry out the directives of the President.").
    On the other hand, we have Saudi Arabia, which, clearly, is not exactly a democracy. It may not be a dictatorship per se since it is a monarchy, but that monarchy rule is absolute ( so it is really a step away from NK ). And you may think they don't need population's consent, but you would be surprised how much consent you need when the population is used to 'harsh but oil is flowing so we get something out of it' trade off.
    In other words, I am not saying you are wrong, but you may want to re-consider your priors.
  3. rurp · 1790117065 · focus · HN ↗
    Great point. Less deadly but just as incompetent was the utter failure to refill the Strategic Petroleum Reserve when gas was cheap for a year leading up to this war of choice; a war that was certain to massively disrupt global energy markets. Energy disruptions had been an obvious consequence to every expert and non-expert who had looked at attacking Iran for decades. Yet the people who chose to start the war managed to be caught completely off guard by that.
    Just like clowns in charge were caught off guard by the fact that Iran has a drone program, resulting in the loss of dozens of air frames and much of our strategic munition stockpiles. Iran literally jump started Russia's drone program in a high profile way, and yet trump/hegseth somehow completely failed to plan for how to counter those drones and had to resort to lobbing scarce multi-million interceptor missiles at them and watching our bases take heavy damage.
    These are just some of the failures we know about. The administration has been extremely tight lipped about these failures and there are surely many more that aren't common knowledge. Who knows how many hundreds of other children have been needlessly killed outside of the school strike.
    mainecoder · 1790190469 · focus · HN ↗
    from the perspective of an oil company when the price of oil goes up they make more money generally even if they sell less volume, it is in their interest that oil does not become to abundant that the price falls it is in their interest even if the economy suffers since oil is essential to many there will not be other option but to fill gas to go to work and they want to make the most money then even if eventually other energy alternatives start being considered since the executive team things about the bonus this year not next year or they year after that THIS year and what can be done THIS QUARTER
  4. stackghost · 1790139795 · focus · HN ↗
    >If you're in a shooting war, the ability to select good military targets faster than your enemy, but not necessarily with zero error, is an advantage.
    I mean, yes and no. Merely selecting the targets is not really an advantage in and of itself. I'm sure Iran could select hundreds of targets in the 48 States at the drop of a hat, but they don't have the air power to hit those targets so the air tasking cycle breaks down.
    The air tasking cycle upon which modern NATO air campaigns are predicated and carried out is a cyclical 6-step process:
    1. objectives not yet met/effects required
    2. target development/selection to deliver those effects/meet those objectives
    3. weaponeering, allocation, prioritization (can a hellfire launched from a drone do this or do we need something more specialized?)
    4. actually write and issue the air tasking orders that cause air sorties to be flown
    5. execution of said ATOs
    6. assessment; GOTO 1
    Normally this is a 72-hour cycle and your ops centre would be running 3+ of these cycles in overlapping fashion so that every 12-24 hours you have one loop reaching the top. AI appears to be in heavy use in the target development step, but there are diminishing returns for speed here. At some point there's just no added benefit for choosing more targets because aircraft can only transit to/from the area so fast. Also the other loops are still running and you're potentially developing targets without knowledge of effects being delivered by the other ATO loops so you're now wasting effort. In other words it's not like a computer where more hertz is better; you might get better results by simply waiting for one of the other cycle's kinetic results that will better inform your own target development stage.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.