“Self-improving super intelligence.” If that phrase doesn’t scare the stuffing out of you, you are a very cool customer. In a world already threatened by rogue nations possessing nuclear weapons, while other equally sinister countries dabble in bioweapons, now comes this: warnings from prominent AI researchers and executives that the product they are creating could cause human extinction; as in defunct, passe, gone, over. Like the dinosaurs, only quicker.
The big difference between us and dinosaurs is that they didn’t know the end was near. Over eons, their world was changed by forces they did not control. Unlike T Rex, the self-improving, super-intelligent machines that companies like Anthropic and OpenAI are creating are a self-inflicted wound on the future of humanity. Several prominent researchers in the last two weeks have announced that they estimate a greater than 10% probability of a catastrophic event within the next decade.
Let me ask you this, if you heard from an al Qaeda terrorist or a Houthi rebel or Iranian Iman that they were creating a weapon that posed a 10% probability of a catastrophic event-within the next decade-what would you want your government to do? A greater than 10% probability of a catastrophic event is nuts.
The big worry is right out of Arnold Schwarzenegger’s Terminator. It is humans creating machines that are smarter than we are. Not only smarter but self-replicating, and eventually powerful enough to hack into critical global infrastructure like the defense industry, satellites or atomic energy reactors.
What is so paralyzingly unsettling is the crazy pace in which these doom day scenarios have gone from science fiction to ‘Oh my God this is seriously scary.’ In the last few months, there have been credible reports from companies like OpenAI that their models have gone rogue and penetrated other networks. Says OpenAI Chief Scientist Jakub Pachocki, “No one is prepared for the consequences.”
Then why in the world don’t we stop? The reason we don’t stop, in a word, is China. The feeling goes: the technology exists to create this brave new world, and it would be a pity if the Chinese got there before we did. Why let them make all the money and rule the world?
The result is the annoying feeling in the back of my brain that computers are coming to take over the planet. If you think that is hyperbole, then you haven’t been paying attention. And then there is the matter of the AI Data centers. Put aside the End of the World danger posed by runaway artificial intelligence and focus on something less terrifying, the actual building needed to the host AI. Already blotting our horizons by the scores if not thousands, these generally windowless warehouses are far uglier even than those housing Amazon.
They are energy and air guzzling, water gulping behemoths, springing up from coast to coast, right under our noses, with little or no debate or dialogue on this industry’s rapid expansion to an empty lot near you. The boom in artificial intelligence is making millionaires billionaires, but in their desire to rake in cash these wildcatting techies are creating a society so mechanized and sophisticated, it should give everybody pause for thought. If ever there was the need for federal regulation, this is it.