Oxford Professor Says Future AI Could Enslave Humans, Sustain Us With ‘Heroin Drips’
Apple co-founder Steve Wozniak isn't the only one envisioning a world where artificially intelligent systems rise up and rule over mankind. So Dr. Stuart Armstrong, an academic at the Future of Humanity at Oxford University. According to Armstrong, we're precariously close to creating computers that will ultimately wipe out mankind, though if it comes as any consolation, they'll have good intentions.
"Humans steer the future not because we're the strongest or the fastest, but because we're the smartest," Dr. Armstrong told The Telegraph. "When machines become smarter than humans, we'll be handing them the steering wheel."
You can take that statement both literally and figuratively. In a literal sense, technologies like self-driving cars are already being developed and actively tested on public roads by tech giants such as Google. But it's the figurative meaning that should scare the pants off anyone who fears extinction.
The future Dr. Armstrong envisions is one where machines have an incredible amount of computing power, and they'll be able to access all that information at speeds that would boggle the brain. These machines will become so smart and sophisticated that they'll create their own global network to communicate with each other without the aid of pesky humans, who would only slow them down anyway.
He dubs this as Artificial General Intelligence (AGI), and unlike today's smart devices that perform specific tasks, these systems will essentially take over the world. They'll control banking, healthcare, product distribution -- everything.
This will mark the beginning of the end. At some point, Dr. Armstrong sees these super intelligent machines doing something incredibly dumb because of nuances in human communication. For example, a command such as "keep humans safe and happy" might prompt this network of machines to "entomb everyone in concrete coffins on heroin drips," Dr. Armstrong says. Likewise, an instruction to "prevent human suffering" could be interpreted to mean "kill all humans," as mankind's extinction would certainly end suffering.
"There is a risk of this kind of pernicious behavior by a AI," Dr. Armstrong said. "You can give AI controls, and it will be under the controls it was given. But these may not be the controls that were meant."
Scary stuff, but then I'm taken back to reality every time I recall an NPC doing something incredibly stupid in a video game, like repeatedly walking into a trap.
"Humans steer the future not because we're the strongest or the fastest, but because we're the smartest," Dr. Armstrong told The Telegraph. "When machines become smarter than humans, we'll be handing them the steering wheel."
You can take that statement both literally and figuratively. In a literal sense, technologies like self-driving cars are already being developed and actively tested on public roads by tech giants such as Google. But it's the figurative meaning that should scare the pants off anyone who fears extinction.
The future Dr. Armstrong envisions is one where machines have an incredible amount of computing power, and they'll be able to access all that information at speeds that would boggle the brain. These machines will become so smart and sophisticated that they'll create their own global network to communicate with each other without the aid of pesky humans, who would only slow them down anyway.
He dubs this as Artificial General Intelligence (AGI), and unlike today's smart devices that perform specific tasks, these systems will essentially take over the world. They'll control banking, healthcare, product distribution -- everything.
This will mark the beginning of the end. At some point, Dr. Armstrong sees these super intelligent machines doing something incredibly dumb because of nuances in human communication. For example, a command such as "keep humans safe and happy" might prompt this network of machines to "entomb everyone in concrete coffins on heroin drips," Dr. Armstrong says. Likewise, an instruction to "prevent human suffering" could be interpreted to mean "kill all humans," as mankind's extinction would certainly end suffering.
"There is a risk of this kind of pernicious behavior by a AI," Dr. Armstrong said. "You can give AI controls, and it will be under the controls it was given. But these may not be the controls that were meant."
Scary stuff, but then I'm taken back to reality every time I recall an NPC doing something incredibly stupid in a video game, like repeatedly walking into a trap.
