'Lavender': The AI machine directing Israel's bombing in Gaza

https://en.wikipedia.org/wiki/Computers_Don%27t_Argue

Intended by who? You don't kill 13,000 children by accident.

The Guardian article makes it clear prior to those denials that those higher-up appear to not to care how accurate it is and appear to be making a conscious choice to accept the fact it is highly flawed on the basis that it might kill some of whom they would legitimately claim as valid targets.

It's clear from the operational details discussed in the article the critical target number is largely number of kills, regardless of whether they are any actual material threat, or not.

Cull predominantly the male population and their family members, not assassinate active threats is the overall impression I got of the Israeli strategy.

I must add that anyone claiming the use of AI and inference models in this way is in anyway justifiable needs to seek help. The claim of 90% accuracy is almost certainly over claiming by over 100%.

Its success was so marked that it was immediately decided in 1893 to move a Tabulator to Ellis Island, to count the ethnics from the source with Hollerith's new technology. Herman Hollerith had great success in his own lifetime, the technology eventually becoming the core of the Computing-Tabulating-Recording Company, otherwise known, a decade later, as International Business Machines.

The establishment of this clear process surrounding race - actual race law - was, believe it or not, pretty novel in Western history. A lot of old-timey race policy - like the relationship between a monarch and the Jews, or what exactly a visiting Muslim could or couldn't do (like sell and buy slaves cough Venice cough) - this race stuff was almost always very, ah, what we'd call "tribal knowledge". A Jew in the Middle Ages could have far greater rights and lifestyle than in later periods, but those rights were completely unpredictable; this was true to greater or lesser extent for many "outsiders" in the early European era. Even in 1900 American innovation in race law - based on "Science!" - was a new thing, and extremely exciting to the enthusiasts of folk movements[2] crisscrossing our entire civilization[3] at the time. One of those was Willy Heidinger, who established Deutsche Hollerith Maschinen Gesellschaft to produce license-built Hollerith machines. World events interceded, however, and the German civil service infrastructure to run a census would not be present until much later . . 1933, in fact, when things would get very spicy indeed in the world of race "science".

And then, of course, cataclysm: the end of the European Order.

On the European continent, a debt to truth was paid. A hundred million dead or maimed, nations wrecked, a whole world - a weltanschauung - burnt down to the foundations - below the foundations. But elsewhere - like in the New World - the lesson was not as stark. And in yet other places the inverse lesson was learned: once you determine a person is not a person, you must brutalize yourself and your population immediately, before the soon-to-be-unpeople realizes that the struggle is existential.

Let's wrap this up.

What 20th century Race Science/Race Law were trying to do was make sense of something as complicated as human culture but using the sciences they understood: 19th century statistics, the physics of iron and steam. Those were the sciences with the capital backing, so - of course! - those were the only science that mattered. Today, we're looking at another complex element of the human experience - human language, human consciousness - and again, we're looking at it through the science that's got the most capital backing it: computation. That's how "text" somehow, incredibly, came to contain "language". Or how "scarcity" was represented by "money" - as if there were any N-dimensional descriptions that could adequately vectorize either of those concepts.

Ultimately, when you really dig yourself into these sorts of artificial - if not downright dishonest - "science-y" establishments, when you start imposing them on the world, you don't break out of them easily. Or without damage. The people making use of your LLM widget do not understand the math - all they know, like the race science of previous centuries - is that it's Science-y. It might as well be wearing a Mitre and Crosier.

[1] What those actions were, is a subject for another post. Probably inside a soon-to-be-flagged topic.

[2] The American example in race law was also very exciting to a certain Mr. Adolf Hitler, as well. You can read all about it in Mein Kampf. Hitler's attitude towards America is really fascinating stuff, but an entirely other subject.

[3] And beyond! Ethnonationalism spread like fire, as colonized peoples realized this could be their big ticket towards peerage in the European age.

We all know (if we stop and think) that a person can be both a teacher and a terrorist.

But according to media here almost every victim except top Hamas brass seems to be referred to by their whatever else they were besides terrorists and the terrorists (or even just soldier) part get hushed down.

So if some guy in a track suit and flip-flops uses an anti tank grenade launcher, discards the empty tube, walks away, and gets lit up, then the next day the Internet is awash with videos of the “IDF murdering a civilian!”

For reference, I think both sides are in the wrong in this conflict, and Israel more than Gaza.

However, the Internet is full of armchair international law experts that are being played like a fiddle by Hamas’ propaganda arm.

Speaking of international laws of combat: no protections apply to non-uniformed combatants pretending to be civilians. None. They can be tortured, executed on the spot, whatever.

If you want protections to apply to you, then wear a uniform or never go anywhere near a gun.

For a source, you can check out the Red Cross document I linked. Specifically, Ctrl+F for "continuous combat function" and read the commentary on recommendation V. The Guidance is considered authoritative in legal circles.

https://www.nationalreview.com/corner/hamas-is-using-unexplo...

You completely lose any credibility with this statement. Civilians can't be "stripped" of protections of international law.

It makes sense. Blowing up a military HQ with a clerk in it makes sense. Blowing up a clerk walking on the sidewalk seems like a wasted effort.

You can come up with all sorts of justifications for anything. At the end of the day, time and time again, over the top escalation usually hurts the stronger party. Asymmetrical warfare doesn’t garner sympathy or military advantages to the stronger party.

https://en.wikipedia.org/wiki/Ismail_Haniyeh

Always interesting to me how western diplomats do not just right out reject bombing of diplomatic buildings, but search for stupid justifications, if it's "others" being bombed and not their team.

Or politicians who don't reject targeting of other politicians' families for killing, when it's the politicians they "don't like" or whatever, and even tacitly support it. Or who don't say a word when a hospital is attacked and hundreds of people murdered in it over several weeks, and ultimately destroyed, but blab something about right to self-defence constantly, or IHL which according to legal experts is used mostly to enable mass murder, not to stop it. Kinda paradoxical for a law that was meant to prevent needless suffering.

It's like all these people have a death wish, because they're setting standards for future wars. And there will be future wars, even in Europe. Anyway, I lost all respect for all the idiot politicians I sadly voted for, who justify day and night the murder of medics, whole families, children, starvation, etc., when it's "the other", and are all up in arms when it's "us". I certainly won't be fighting for any of them, when the war comes here. They have 0 standards. I'll let them die according to their wishes and standards.

I give you a gun and say "shoot whoever you choose with this gun, the choice is yours".

I give you an AI powered gun and say "use this however you choose, I have programmed it to automatically shoot in certain circumstances".

In the latter case, I have some responsibility, because I shared in the decision making by programming the gun. Through my code I have put my proverbial finger on the trigger, right next to your finger.

I know roughly ~1000 people. Maybe 10 of them have the physical capability of killing someone, in case you don't know, it's not actually that easy to do it yourself.

Of those not all could mentally do it under anything but the most extreme of circumstances. 2, maybe 3 might be actually capable of ending a life under extreme circumstances.

With a gun probably, at a guess, ~400 - 700 could kill someone if they got anxious/scared enough is my bet. Even if I'm way off it's a lot more than without a gun. Couple of hundred at least. Not 2, or 3!

So yes, I'm sorry, guns definitely, 100% kill people.

And more people will absolutely kill someone if they possess a gun, than if the didn't. And by extension same is true if AI.

I'm interested how you even come up with that response? It's obviously factually and logically wrong. What makes you think it makes a reasonable argument to anyone?

Also, worth pointing out, thar AI in this case is insanely unfit for it's purpose (unlike a gun) and will have randomly killed lots of innocent people, even if the AI algorithm says otherwise.

The computer basically appeared to randomise a high number of people to kill based on a very shallow dataset, weak data linking and a high desire to kill people who are Palestinian.

It's hard to read the details from the Guardian article and think of it as anything other than a randomised Israeli state murder machine. I can't envisage it being accurate to a point any reasonable person would accept it be used against someone they know, in any circumstances.

That it targeted 10's of thousands is utterly horrifying. That it was involved in any actual battleground scenario makes me think those involved in it's creation and sale are culpable.

Obviously, nobody in an international court will be able to say "... but the AI did it!" - This is just far too easy as a way out. There are rules to AI Usage and one of them is not a usage like that - as already said somewhere else: The AI is only as ethical/moral as the humans behind them.

Why would Hamas use anything other than clearly uniformed soldiers, marked military vehicles, and civilian distanced military installations?

I will try to solve your doubt. Gaza is a closed area. You can't just cross freely the frontier unless Israel allows it. Before that, humanitarian organisations are required to inform directly to Israeli army about everything that they want to do, where they want to go, how many, an when. They are clearly identified by a logo at all times. Israel had every single piece of info necessary to avoid bombing aid workers and had it for weeks

Despite that, Israel can't avoid to keep killing this workers; and they were doing the same repeatedly, systematically, for the last six months, in multiple attacks that last tens of minutes (maybe even hours?), while singing "oops!, I did it again".

The theory of the honest mistake is getting really difficult to swallow.

It is. The IDF shot once, then when a vehicle came to rescue survivors, they shot again. Then the third attempt at rescueing survivors, they shot a third time. Purposefully murdering aid workers that coordinated with the IDF before entering the area.

https://www.msn.com/en-us/news/world/idf-strike-on-gaza-aid-...

“…were traveling in a convoy that had been coordinated with the Israel Defense Forces and was following an IDF-approved route. The vehicles had GPS trackers and SOS beacons broadcasting their positions”

> One source stated that human personnel often served only as a “rubber stamp” for the machine’s decisions, adding that, normally, they would personally devote only about “20 seconds” to each target before authorizing a bombing — just to make sure the Lavender-marked target is male.

> According to the sources, the army knew that the minimal human supervision in place would not discover these faults.

I took this to mean that a human did press the "approve" button on the computer's recommendation. Though they make clear they were basically "rubber stamping" the machine recommendation.

But to my point:

> “There was no ‘zero-error’ policy. Mistakes were treated statistically,” said a source who used Lavender.

What is the "zero-error" alternative approach for dropping bombs in a war, or firing rockets for that matter? I don't understand the implicit comparison between this approach to targeting and a hypothetical approach that allows war to be waged without any innocents dying or buildings being destroyed. This system should be compared to whatever the real alternative is when it comes to target selection. Again I know nothing about military strategy, I'm hoping someone with more experience will speak up.

To use an analogy: if we are talking about self-driving cars, the rates of collision or death should be compared the rates of collision or death in cars driven by humans. Comparing against some imaginary scenario where cars have no collisions and cause no deaths doesn't make sense.

"Hamas terrorist" criteria: a male of fighting age, give higher weight to those congregating with others of fighting age. Basically take out a generation of Palestinian men and you're all set. Lovely.

>This sentence is horrifically dystopian... "in order to save time and enable the mass production of human targets without hindrances"

Reminds me of similar industrial thinking of a certain previous fascist government.

They just spout a high number that is not 100% (clearly civilians are being killed publicly undeniably ) claiming 100% would be too obviously ridiculous.

More than half of 32,000+ (more under the rubble) killed are woman and children, Hamas is still quite able to fight, hardly any hostages has been recovered .

Israel labels any sort of civilian organization as hamas including journalists, medical and aid staff. 200 UN staff and 100 journalists are dead so far . Israel’s argument is UNWRA terrorist aiding and journalists were also secretly Hamas and doing non journalistic stuff when killed so they include them in legitimate targets .

If you consider everyone is Hamas unless otherwise proven then 90% is possible .

There is no realistic way an algorithm was designed factoring in the level of destruction of infrastructure never seen in any real world data and also benchmarked accurately.

That said, any organization might do something if it’s 90% accurate. Assuming it even is (doubt it), I think any fair evaluation of such a technology must ask:

What is the accuracy of inexperienced humans in the same position who are rushing through the review during a blitz invasion? If they have battle experience, what about them, too? (I’m assuming most won’t.)

Is the system better than those humans or worse? How often?

Do the strengths and weaknesses of the system allow confidence scores on predictions to know which need more review? Can we also increase reviews when the number of deaths will be high?

That’s how I’d start a review of this tech. If anyone is building military AI, I also ask that you please include methods to highlight likely corner cases or high-stakes situations. Then, someone’s human instincts might kick in where they spot and resolve a problem even in the heat of war.

I don’t understand how to come to this. War is crap, not a dinner party. There’s always a human on both sides who will drop a bomb and laugh on camera, with no responsibility. Go watch it (actually don’t, it’s NSFL). Reading this thread feels like everyone watched and believed in that movie where they tried to select and eliminate a target for 2 hours with futuristic hi-tech. A human hesitates to press the button before the war. When in it, he will only be concerned with things like ammunition saving and tactical nuances. There’s not much more morals in a human who usually sits there at the button than in AI automation.

We know that humans can make mistakes, due to a multitude of reasons. They can be tired, moody, distracted, stressed out, time-pressured, simply not care enough, etc, all contributing to making the wrong call. But a computer does not suffer from such issues. Secondly, a computer (program) is able to perform billions and billions of computations within some time period in order to ENSURE that doing this thing with grave consequences is absolutely warranted.

Maybe for some domains we can tolerate errors from AI, but when deciding whether a person (and everyone around them) lives or dies, surely simply being on average even more accurate than a human is not enough. "Killbots" MUST be extremely heavily regulated.

Reminds me of that story from probably 5-7y ago. Someone wanted to use AI to classify photos of tanks as soviet vs US. So he went to a US tank museum and took lots of pictures of the tanks under every angle. Did the same in a soviet tank museum. The resulting model worked great on that training dataset. Then he tried on photos outside of the training dataset. Turned out that it was cloudy the day he visited the US museum and sunny for the soviet museum, and the model used the color of the sky to classify.

Having a human to make those decisions is better because this human can be judged if commits war crimes or genocide or violates international war laws.

A computer can't be jailed and this is the real power of designing this system. To hide the criminals on a black box so nobody can be made responsible

If this is about the hostages, why hasn't Israel agreed to more ceasefires to save them?

The only thing that made this time a bit different is the crazy, almost hard to believe, switch from the Ukrainian conflict and how it was seen and portrayed... To western countries staying completely silent when again, it's our side doing it. Well it wasn't hard to believe but it just made it a lot more blatant.

Israel doesn't really care though since israeli officers routinely go on public tirades that amount to mask-off allusions to genocide ("wipe Gaza" "level the city to the ground" "make it unliveable"), with again 0 consequences at all. Even Russia at least tries to not have Russian military officers just say the quiet part out loud.

Maybe it is. Maybe it isn't.

Some questions worth asking: what is international law? How is international order maintained?

I agree that images and footage from Gaza are disturbing. But I encourage you to think systematically about what it is we are seeing.

When the US was in Afghanistan, Al Qaeda learned that the US (generally) won't shoot ambulances. So what became the most valuable vehicle to Al Qaeda? Hamas took notes, but Israel doesn't seem to care as much as the US.

Also, besides all that, once something is used for military operations, it is fair game as a military target. Regardless of civilians. When the law was written it was assumed that governments wouldn't intentionally use their civilians as protection.

https://archive.is/2024.04.02-205352/https://www.haaretz.com...

Yet still, even that its the most important, friendly fire still happens

> With unintended strikes, there's "we work hard to avoid this, but based on bad intel made a rare, tragic error," and "we've encouraged RoE that foreseeably makes tragic errors frequent, but this looks bad and in hindsight wish we hadn't done it."

> Israel's strike on WCK food aid workers is the latter

Israel has long had pretty plain issues with its rules of engagement. Recall that earlier in this conflict, the IDF shot three of the hostages whose recovery is one of the main goals of the operation!

On the other hand of course there are also those that jump on any claim that makes Israel look bad. Claims of which there are many. Of which far too many have become pretty evident. Which far too many people do not want to be true and will ignore.

So what can one do? I guess keep an open mind and give claims a couple of days to be proven or disproven. Only then judge.

Readers might still find it helpful to read both pieces, of course.

It seems to me that Israel's overall position - politically, diplomatically, and in terms of physical security - has become much worse since the October 7 attack and it has been their own actions that are responsible for the change. A different response should have seen them politically and diplomatically strengthened.

I understand the emotive reasons for not doing so, but I think most people would consider that Israel has bungled their response to October 7.

I would call this attack on Gaza a total failure. If nothing else a failure of humanity.

It's looking more and more like the 'winners' in this situation are Hamas and the losers are the Israeli government, the US government, and the Israeli and Palestinian people.

Months ago the IDF claims that Hamas is operating out of tunnels under the hospital, they raid the hospital. A few weeks later, they raid the hospital again, and again, and again, up until this recent raid some days ago.

You claim Hamas is diminished, but how diminished can they really be if they keep popping up in the same predictable place over and over and over? In North Gaza, the place the IDF has been fighting to secure for the longest, Hamas just pops up one day in the same place they have many times before and so they have to raid the hospital again.

This is all according to Israel's own claims. If Israel's claims are true, then Hamas is one of those mystical enemies that pops up everywhere and is super strong and justifies all sorts of things to defeat, while also being weak and "diminished".

I think the truth is that Israel wants to destroy the health system in Gaza and drive the people out. Shifa hospital is basically rubble at this point.

"we really need to missile this guy or he will kill more" vs "well we got 37 badies and also kim and yashonda, damn i really liked yashonda"

Actually after writing this my mind went farther, "since yashonda was a good person we actually have a whole bunch of hard facts about how good a person she actually was, did a lot of help for her community and was a real pillar of helping the next generation of kids be less violent...too bad we didn't add any of that info into the kill-algorithm "

I wonder if you would say the same on the other side where every male or female above 18 years is required to serve in thr military and in the reserve afterwards? [1]

By your argument would you say that all of these are legitimate targets?

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

I also don't think there is a way to complain about abusing flags other than emailing the mods; I have no clue about the effectiveness of this complaint.

It's also currently dropping rank on the front page, despite being heavily upvoted.

Admins can, and do, prune entire branches of comments off of posts.

These two methods would take a bit more work than just banishing the topic entirely, but with topics like the first time that "AI" kill lists are publicized, maybe exceptions should be made.

For a high quality piece of tech-related investigative journalism like this, flagging is simply censorship.

Considering what regularly doesn't get flagged on this site related to AI, conflict, etc., this topic seems to fit in.

I guess all you have to do, if you want to suppress information about something, is to ensure that its comments always devolve into unproductive discussions. Funny, I once read about this as a tactic for controlling information flow in online communities...

im not sure what other tools exist other than a block button like X

I wish people would let people decide for themselves what is productive or not...

Kind of related thought - is there a topic you think is more divisive? And also, is there some way that this is measured officially or unofficially?

Having a human in the loop prevents bad-faith actors from abusing the system to suppress information and discussions.

> According to six Israeli intelligence officers

Not 1 reservist, or 2 retired officers, or 3 contractors, but 6 active serviceman - whose day to day job is to figure out how to hide secrets.

There are more statistically impossible statements.

Development of tools of death is not a good guy/bad guy thing. The "bad guys" think the "good guys" are bad.

I think "killing" is bad, no matter who develops the tools.

Edit: And the humans who approved the list should be help accountable, of course.

Western mainstream media has been very passive when covering the current situation in gaza, especially when you contrast it with how they covered the war in ukraine just 2 yrs ago. Its just that social media has allowed people to break through the canned media narratives.

WW2 was a considerably different war in scope, origin, and patterns of escalation.

The use of napalm in Vietnam triggered widespread protests.

They did and the newspaper coverage is the main reason why the Vietnam war stopped.

There is also another AI system that tracks when these target get home.

Additional automated systems, including one called “Where’s Daddy?” also revealed here for the first time, were used specifically to track the targeted individuals and carry out bombings when they had entered their family’s residences.

I think "assassination" colloquially means to pinpoint and kill one individual target. I don't mean to say you are implying this, but I do want to make it clear to other readers that according to the article, they are going for max collateral damage, in terms of human life and infrastructure.

“The only question was, is it possible to attack the building in terms of collateral damage? Because we usually carried out the attacks with dumb bombs, and that meant literally destroying the whole house on top of its occupants. But even if an attack is averted, you don’t care — you immediately move on to the next target. Because of the system, the targets never end. You have another 36,000 waiting.”