It’s a vision of a possible future. Tiny, friendly-looking drones fly out of a truck and buzz across a college campus. There they work their individual ways into a classroom and explode against several students, killing them.
That’s a scene from “Slaughterbots”, a short video from the Future of Life Institute outlining one possible direction for drones and related technologies. It is chilling, provocative, and well done. Watch:
Let me comment on the video in detail, then look into this as a futures document, followed by some more ideas.
0:05 – a good representation (and maybe parody) of the TED/Steve Jobs presentation style. Bonus points for the techbro CEO/spokesperson.
0:55 – added to the tiny drone are facial recognition and sensors.
1:15 – throws killbot into the audience. Cute move.
1:35 – note the anti-human stance, the love of command.
2:06 – boosterism: “they cannot be stopped”. No mention of counterforce. I’ll get to that below.
2:22 – “thinking big”: more Silicon Valley language, supporting the key feature of scaling up.
2:50 – we cut jarringly away from the TED talk to a series of news video clips.
2:55 – “we have a distribution network” is spoken over images of other distribution networks.
3:03 – it’s barely there for less than a second, but you can glimpse a group of slaughterbots forming a giant swastika.
3:04 – U.S. Senators assassinated.
3:29 – key feature: no signature on a slaughterbot attack.
3:46 – slaughterbot arms race.
3:56 – short story about bots killing university students in Scotland, presumably over a protest.
5:00 – claim that personal firearms can’t stop the bots. People encouraged to stay indoors – in general, or in response to an attack?
5:04 – slaughterbots named by a British-accented tv news announcer.
5:12 – the attackers, who launch a small swarm from a van. They are white men, and otherwise unmarked.
6:07 – the higher education attack is apparently worldwide, with nearly 9,000 students killed.
6:21 – some commentator argues that slaughterbots are chilling freedom of expression and dissent.
6:40 – back to the lethal techbro, who emphasizes data and social media. “You can target an evil ideology.”
7:09 – the fiction ends, and one of the creators, Stewart Russell, a Berkeley computer scientist, reflects on what we’ve just watched. He emphasizes AI and giving machines too much autonomy.
7:42 – directs us to a supporting website, http://autonomousweapons.org/ . The site seems to prefer the term “killer robots”.
So how does this fare as a futures document? Pretty well. It quickly portrays a near-future scenario, fleshing it out with visual and oral storytelling. It balances technical discussion with strong emotions (a mother watching her son killed on live video). It steps back at the end to help frame reactions. Tonally it aims to instill fear and reaction, perhaps embodying my aphorism: the best way to predict the future is to prevent it.
I naturally have questions, and want to push the story even further ahead. I wonder how these slaughterbots would play out.
To begin with, there are technical obstacles to realizing their existence. As Lawfare points out, it isn’t cheap to build assassin drones that small right now. There are limitations on battery life; a slaughterbot controller would have to deploy the things relatively close by. And facial recognition is better, but still flawed. So these critters aren’t tools of the present, but of the near to medium term future.
Moreover, while the video’s frantic tv talking heads shriek that there’s no stopping them, defeating drones is a growth industry today. People are trying out guns, lasers, and even predatory birds. Drone versus drone combat is emerging as a new combat arena. And there’s always the hacking option.
As for anonymous killerbot attacks, we have to consider the wide world of surveillance. We can assume not only growing state and corporate data- and media-gathering, but also individuals using mobile devices to record anything that moves. Imagine a state security apparatus asking Google for help in tracking down telling searches, or victims’ families crowdsourcing popular surveillance.
So why does the film’s main example take place in universities? The killing could be the act of a repressive regime or international organization opposed to some rising student movement. Perhaps it represents hostility to anti-racist activism or protests against rising fees. Is the video’s use of a campus designed to suggest slaughterbots will repress learning, or because for some school represents a safe space in some sense?
But let’s take this a bit further. How else might these “slaughterbots” play out in practice? What kind of futures could unfold?
Killbot versus killbot People and organizations, terrified of the whirring assassins, could stock up on defenses. Those might include: anti-slaughterbots, designed to seek out popular designs and block or destroy them; face disguises (masks, hoods, appliances); area denial tools (metal nets, sonic generators); bats and sticks. Some could set up versions of panic rooms, with surfaces too think for the little killers to penetrate, or surrounded by Faraday cages or limited EMP generators.
Debris Once prices fall and these scale up in numbers, would-be killers may have to deploy larger and larger swarms to get past defenses. Imagine thousands whirring and whapping at each other in a teeming metallic cloud. Parts will break off, and individual bots plummet to the ground. The aftermath will be a pile or field of metal and plastic bits, giving rise to professional disposal companies and popular irritation. It might resemble a coarsely grained version of Neal Stephenson’s idea of post-nano toner. (HT Rob Henderson)
Hidden and delayed killbots The video shows aircraft and vans dumping slaughterbots for immediate and fairly visible use. What happens when a bot-minder decides to get stealthy? Imagine concealing bots amid foliage in a rural area, or in junk in a city. They could stay hidden for hours or days, if they manage power correctly, only to spring out on a timer, pre-arranged signal, or analysis of the immediate situation. Like mines after 20th century wars, killbots could last for a while, lashing out after the mission’s timeline has come and gone.
Suicide by slaughterbot Suicide is an act of violence most societies refrain from examining, but it remains a significant part of human existence. When will the first person enter their features into killbot software, and wait for the end delivered by airborne explosive? Will some attempt a bot version of suicide by cop, deliberately egging on a bot-owning force (police, gang, etc) to goad an attack on themselves, perhaps donning a face mask of a known criminal?
Stories These killbots would have a fantastic presence in stories of all kinds. At some point people will imitate the snapping explosive sound in conversation, and all kinds of media (music, games, movies, VR) reproduce or sample the sound for dramatic effect. Kids will tell each other stories about slaughterbots, who will then enter urban legend (the possessed bot! the invincible bot). Slaughterbots will enter language, acquiring nicknames, becoming nicknames for gangsters and sports figures, playing a role in proverbs (“irresistible as a slaughterbot”).
At school At some points schools of all levels will have to ban students from building slaughterbots.
Meanwhile, campuses will have to choose to either ban or support faculty research into the technology. It’s a new twist on an old problem.
In policy Governments will scramble to develop and promulgate laws about killbots. Some will seek to ban them outright, while others make room for their own implementations. Naturally an arms race will occur, with states and companies vying for advantage, influence, and market share.
That’s enough for now (bzzzzz). Do you think (bzzZZZZZ) slaughterbots are (BBzzzzZZZZ) possible? If so, (BBBZZZZZ) where might they go nexBANG
(thanks to friends on Twitter and Facebook for conversations about this video)