A Drone Killed Three Ukrainians. It Was Guided by A.I
nytimes.com
nytimes.com
OTOH, a bit of perspective might be useful. Compared to historical large-scale European land wars since the relative "perfection" of modern artillery (c. 1900), the civilian casualties we're seeing in modern Ukraine are - while still horrific - quite low by comparison. The Feb'45 air raids on Dresden killed >20,000 civilians in 3 days.
Bluntly put: "A.I." really ain't the problem, and even a magically perfect A.I. ban would not meaningfully reduce the horrors of war.
This is scary because the more we remove humans from the loop, the more power we give to fewer decision makers. Recall all the times during the cold war where missile launches were detected and the crisis was averted by deliberate decision making on the part of lower level officials, such as Stanislav Petrov. Then there is the more sci-fi risk of fully autonomous systems that run out of control.
Regarding the scale of casualties in Ukraine, could one view this conflict similarly to the Spanish Civil War, as a staging ground for powers to try out technologies that may be used in a much larger war later?
Sounds very similar to artillery shells, grenade-and-wire booby traps, homing torpedoes, land mines, ship-counting naval mines, spring guns, and such - humans are closely involved with [launching|placing|setting up] each one, but victim selection is pretty random. All of those were widely used in WWII, or earlier.
The article's "model airplane" drone did not build itself. It was not built by an autonomous "robot factory". It was not part of a massive wave launched at Skynet's orders. Compared to even an early 1990's Tomahawk missile - https://en.wikipedia.org/wiki/Tomahawk_missile - its capabilities were meager.
The problem in the cold war was not rogue robots, but human command structures having to make yes/no decisions on mass launches of non-recallable nuclear weapons under extreme time pressure. The humans hated that situation - but felt that the possibility of a preemptive nuclear first strike left them with no better alternatives.
Vs. the drones we're seeing in Ukraine have nothing resembling a national-scale preemptive first strike capability, for that to apply.
Even during war, people follow some laws. I don't think terrorists would care if exploding a nuclear reactor could kill thousands
It's been true for decades that a lone agent with a shoulder-launched missile - https://en.wikipedia.org/wiki/Man-portable_air-defense_syste... - could shoot down a big airliner either landing or taking off at almost any civilian airport. Yet that pretty much doesn't seem to happen.
Similar for a great many ways in which a motivated-enough crazy could create a mass casualty event, with no need for modern drones or AI. Might you recall the https://en.wikipedia.org/wiki/Oklahoma_city_bombing ?
(America's 3-letter spooks might claim that they have been secretly foiling all the evil plots - but their Creep State has had so many public failings that it's extremely hard to believe.)
Even Mr. bin Laden's 9/11 attacks aimed at highly symbolic targets - World Trade Center, Pentagon, Capital Building(?). Not at (say) football stadiums packed with hundreds of thousands of ordinary people.
SO - if you work in (or adjacent to) the physical security field, then I'd actually worry about a bit violent crazies with drone or AI drone tech. Similar if you're in a war zone, or some place where the violence is real & regular. And similar if you write fiction, and want to make your Day of the Dronez novel a bestseller.
Otherwise - It's okay if you have an active imagination. Or prefer worrying about drone/AI tech to worrying about (say) your grandmother's failing health. But the things really worth worrying about are heart disease, cancer, accidental injury, stroke, ..., lack of successful relationships, ..., alcoholism, ..., homelessness, ...
Or about one Mertens, Frankfurter Allgemeine journalist, asking the jewish fuhrer of Ukraine in Belgrade "how can we help you kill more Russians?"?
..thought so..