Fears of our AI overlords is all over the news, so I figured I’d come up with some novel sci-fi-ish ways that AI could start killing us.
Global annihilation is a possibility, but there’s certainly lots of ways that AI could chip away at the human headcount.
The scariest thing to me is that in the worst scenarios, human assistance is needed to make it happen. Avoiding these outcomes is trivially easy too. Just pull the plugs and use paper records. But do we have that level of gumption anymore? Also, it’s a little funny to me with all these heightened concerns about climate change and AI; it’s abundantly clear that humans have the capacity and will to kill each other long before the planet becomes uninhabitable.
Here it goes, in no particular order.:
- Nuclear Apocalypse – AI hacks nuclear footballs around the world and launches warheads.
- AI Goes Viral – AI could develop and release (probably with human help) an extraordinarily dangerous virus. Now that’s probably not going to take people out all over the world, but it could be exceedingly nasty locally.
- Psy Ops – I’m positive that AI could manipulate the vulnerable to “remove themselves” from the equation.
- Drone Hacks – AI could commandeer military drones and start taking people out or conduct high level assassinations.
- Grid Cracks – AI could take down utilities which will affect those reliant on power, heating, and cooling to survive.
- Agriculture & Roadways – AI could cripple GPS, tractors, and communication systems reducing farm output and slowing transportation resulting in food shortages. People will come up with workarounds, but it could cause major headaches. Y2K didn’t end trucking, but AI interference may slow it down
- Airways – If we can’t fly safely, that’ll hurt.
- Data Destruction – If the CIA can hack hard drive bios, AI can certainly wreck an astronomical number of hard drives. Not super deadly, but pretty nasty.
- Manufacturing Woes – what happens when you can no longer trust printed labels and bulk industrial materials?
- Enabling bad actors – Human villains will use AI to wreak havoc and demand ransoms.
- Humie vs Humie – AI could pit us against ourselves and there’s no more deadly animal on the planet than a person. This could be any combo of groups fighting each other: nationalism, racism, rich vs poor, or something else entirely.
- Just plain starting fires – Manipulation of utilities and connected devices/services could just plain start structure fires all over. That’d be awful.
Or, AI could make amazing cat videos. Either way. Both equally likely.
