Hacker Cinema
Slaughterbots
2017 · Film · short · 8 min
Directed by Stewart Sugg
A UC Berkeley computer science professor helped to create a video that imagined a world where nuclear weapons were replaced by swarms of autonomous tiny drones that could kill half a city and are virtually unstoppable. Stuart Russell, the professor, said these drones are already a reality.
9 positive1 negative6 neutralfrom 16 commenters
Where to watch
No streaming data for this title.
Videos commenters linked
From YouTube links in the comments below. Most are clips and trailers; “Full” means the uploader says so.
16 comments on Hacker News
I can think of several possible messy outcomes that would be able to directly affect me, not all mutually exclusive:
- Job loss by me being replaced by an AI or by somebody using an AI. Or by an AI using an AI.
- Resulting societal instability once blue collar jobs get fully automated at scale, and there is no plan in place to replace this loss of peoples' livelihoods.
- People turning to AI models instead of friends for emotional support, loss of human connection.
- Erosion of democracy by making authoritarianism and control very scalable, broad in-detail population surveillance and automated investigation using LLMs that was previously bounded by manpower.
- Autonomous weapons, "Slaughterbots" as in the short film from 2017
- Biorisk through dangerous biological capabilities that enable a smaller team of less skilled terrorists to use a jailbroken LLM to create something dangerous.
- Other powers in the world deciding that this technology is too powerful in the hands of the US, or too dangerous to be built at all and has to be stopped by all means.
- Loss of/Voluntary ceding of control over something much smarter than us. "If Anyone Build It, Everyone Dies"
The film 'Slaughterbots' presents a scenario which could be either of those, but is implied to be the latter.
So the drones in the Slaughterbots short film were depicted to be way too slow.
Obligatory link to the short film (future documentary) “Slaughterbots” (2017), which depicts exactly this in harrowing detail:
https://www.youtube.com/watch?v=O-2tpwW0kmU
It's a short movie from Future of Life institute, a non profit researching Ai. The movie is of course over the top but it catches really well that "the algorithm says so, it must be right"
Another short movie I linked from the same channel is Slaughterbots: if human: kill()
https://news.ycombinator.com/item?id=42511342
Sounds like you also watched the short movie "Slaughterbots".
“Only a net can help. And I predict that soon we will have to put up such nets above our cities, or at least government buildings, all over Europe.”
That’s quite some food for thought, especially in a non-war context.
Reminds me about the short film “Slaughterbots” [1] and that one Black Mirror episode with the weaponized robotic bees.
[1] https://www.youtube.com/watch?v=O-2tpwW0kmU
See “Slaughterbots” short film: https://youtu.be/O-2tpwW0kmU
Wikipedia has additional context and analysis on the short feature and its impacts: <https://en.wikipedia.org/wiki/Slaughterbots>
The project's website is at <https://autonomousweapons.org>
And subsequent developments in Syria, Iraq, Afghanistan, and Ukraine since the film was released in 2017 show how the concepts illustrated are being realised in real-world conflicts, highlighting both capabilities and limitations of the concept, though I'd strongly suggest that the full potential has yet to be realised.
One interesting distinction is between the nature of precisely targetable weapons as with the slaughterbots, and stochastic ranged weapons, which date back to at least WWII and the long-range but only loosely-targeted attacks via V-1 "buzz bomb" and V-2 rocket bombs, or more recently through cheap and inaccurate rockets such as the Qassam used heavily by Hamas (see: <https://www.nytimes.com/2021/05/13/world/middleeast/gaza-rockets-hamas-israel.html>), or the Russian S-8KL (see: <https://www.newsweek.com/russia-cluster-bombs-dumb-rocket-s8kl-rockets-1820504>).
Precision is useful against specific military targets or civilian leadership. Stochastic attacks have long been used against general populations or forces, though the morale effects may well backfire unless the lethality is truly overwhelming.
The questions raised in the video about countermeasures and the consequences for use-of-force are interesting and I'd very much like to see further discussion of those topics if anyone is aware of any.
I remember the first time I learned about drone swarms was this YouTube video called Slaughterbots: https://m.youtube.com/watch?v=9CO6M2HsoIA
Of course at this point, my country had been using drones to kill people since I was a small child, which I think is ironic.
In the Slaughterbots scenario, each specific drone has sensors and AI good enough it can pilot itself and doesn’t need an external connection that can be detected or disrupted and is good enough to target individual people with facial recognition. I believe (this is not an area of expertise for me) that the drones depicted in that video are still far more advanced than what currently exists currently (but I think it could probably get there within the next 10-20 years), whereas the video linked shows a more realistic drone swarm attack where it’s targeting infrastructure and the blast radius is larger than being able to target an individual's head, but still smaller than traditional UAVs.
It’s an upsetting watch, but the part of me that enjoys speculative science fiction (I love evaluating how realistic the apocalypse is in zombie movies) always found the attempted countermeasures employed to be silly and unrealistic. Covering windows and leaving only tiny circles so no drone can get through etc…
But at the same time, the video struck a chord because as soon as I saw it, I knew some form of this was coming and likely within the next decade. But mitigations and countermeasures are being developed in tandem. I think that's the major thing the Slaughterbots video got wrong.
And there's many different potential countermeasures, they don't all have to be 100% effective, but together they form a kind of "net" that will make these attacks a lot harder to pull off.
For example, they now have a portable machine that can disrupt a drone swarm's sensors, which will at the very least require additional shielding (and increase cost per unit) which removes some of the benefits. It's possible the technology will prove even more fruitful after additional development and make drone swarms mostly untenable for warfare except in rare situations.
This is a very direct countermeasure, but I could also imagine other indirect measures like regulating consumer drone manufacturers and forcing them to add/remove features that make it more difficult to turn them into a weapon. The same way printers can't just print currency and GPS won't work if it's moving too fast.
Ultimately, for war it just makes it another type of weapon/tactic and there's already infrastructure for controlling how those things move between countries and intelligence groups that work to keep them away from terrorists etc...
Which leaves the possibility of a sole actor or small group without state backing turning consumer drones into these swarms. Which could happen, it could be a new vector of terrorist attack. But they'd still have to source explosives which is already monitored, etc... In the Slaughterbots video, these swarm attacks are extremely efficient and effective, killing 5000+ people at a time. I think a similar attempt from a homegrown terror cell would not be more effective than a mass shooting or traditional bombing in terms of casualties (< 50.) I think for it to become more effective, it would have to be done many times and iterated on by the same group, building up institutional knowledge. It's just too complex.
Like the idea of a "dirty bomb", I think it’s one of those things that is theoretically extremely dangerous but due to a mix of factors is unlikely to result in a situation like the one depicted in that video. A matrix of countermeasures will just make the juice not worth the squeeze.
I think about it like guns. There is a lot of gun violence in the United States, but it stems more from ideological gridlock than the government not being able to stop it if they had the will to.
The one thing that does worry me is the possibility of targeting individuals. So if you have a tiny drone that can explode and isn’t part of a larger swarm, then you could potentially murder someone in a way that is very hard to investigate. Much harder than a homicide involving a gun where someone had to pull the trigger. It could be the next evolution of SWATing someone.
Sci-Fi Short Film “Slaughterbots” | DUST https://www.youtube.com/watch?v=O-2tpwW0kmU
Killer bees are in S02E06 "Hated in the Nation", https://en.wikipedia.org/wiki/Hated_in_the_Nation
There is another one with killer dogs in an apocalyptic wasteland (in black and white), S04E05 "Metalhead".
The short film "Slaughterbots" (https://www.youtube.com/watch?v=HipTO_7mUOw) fits it even more precisely though, I'd say.
Must watch - Slaughterbots
https://www.youtube.com/watch?v=9CO6M2HsoIA
The substantive part of this conversation comes about 35 minutes in, when Allen describes the DoD's five AI ethics principles: that any deployed AI systems must be "Responsible, Equitable, Traceable, Reliable, and Governable." [1].
I think "Responsible" raises some hard questions. Today, any weapons use by a robot might require a responsible human operator to first (remotely) pull the trigger. But what if an adversary allows their robots to fire without waiting for human permission? The first side to do so would gain a critical advantage, and we cannot have an autonomy gap.
Some have raised conerns that in the extreme case, this could lead to a world of machines anonymously killing without attribution. If you haven't, see Stuart Russell's compelling short film "Slaughterbots" [2].
[1] https://media.defense.gov/2019/Oct/31/2002204459/-1/-1/0/DIB_AI_PRINCIPLES_SUPPORTING_DOCUMENT.PDF
[2] https://www.youtube.com/watch?v=9CO6M2HsoIA
I'm surprised there's no mention of the short film "Slaughterbots" on this exact topic:
https://www.youtube.com/watch?v=HipTO_7mUOw&vl=en
It's very well done and has an HN vibe to it (with a dark twist, of course).
No fly zones will only keep good guys out. The same is true of any regulation-based solution. Anybody can build a drone at home that ignores no-fly zones and doesn't broadcast any serial numbers, and then operate it remotely, with very little chance of getting caught.
I don't know what the answer is, other than to hope that most people continue to be mostly good.
In case you haven't seen it, the short film "Slaughterbots" is worth a watch: https://m.youtube.com/watch?v=9CO6M2HsoIA