For most people, artificial intelligence is Siri, the lovable but modest digital assistant on iPhones.
She can’t really do much beyond answering simple questions. She’s harmless.
Advanced artificial intelligence is far from that. Algorithms are constantly learning, often in unusual ways, and at an exponential rate. It’s one of the cornerstones of our age.
Co-founder of Google DeepMind Mustafa Suleyman attends a Q&A with Special Projects Editor forTechCrunch, Jordan Crook during day 1 of TechCrunch Disrupt London at the Copper Box on December 5, 2016 in London, England. (Photo by John Phillips/Getty Images for TechCrunch)
Unfortunately, nobody really knows how they learn. And that should scare you, a lot.
When Siri doesn’t know the answer to a question, she demurs. She may quip or make a snarky comment to obfuscate, then send you to Wikipedia.
“Hey Siri, what is the meaning of life?” Siri: “It’s nothing Nietzsche couldn’t teach you.” Cute.
When Google wanted to improve the way its machine-learning software translates languages, it started by feeding it massive amounts of data. In short order, the software figured out common translations. Soon, accuracy rates improved. Later, the software began translating between languages it had not even studied.
Initially, engineers were perplexed. They later concluded the AI had devised a brand-new language, or “interlingua”, to make sense of language pairs. Once it had the cipher, the rest was a snap.
When Nvidia decided to enter the self-driving car race, it started with neural networks watching humans drive, sometimes badly. Soon, the AI could steer, brake and accelerate just like a careful, confident, skillful person. Later, using a batch of sensors and a learning, trunk-mounted brain, the AI could drive well enough to navigate winding dirt roads, driving rain and thick fog.
Engineers admit, the system is so complicated even they don’t really know how the software is arriving at all of its decisions, MIT Technology Review reports.
This is the part where most people start worrying about Hollywood-style Terminators.
I will admit, the idea of machines making their own choices based on processes that even their creators don’t understand is unsettling, at best. And plenty of smart people like Stephen Hawking, Bill Gates and Elon Musk have sounded alarms about unintended consequences.
In a world where machines learn, why wouldn’t they develop emotions and all of the other frailties that cause humans to act irrationally? It could get seriously weird, and dangerous.
That does not mean there will not be terrific investment opportunities.
AI is a game-changing technology. It is already having a positive impact on the bottom line for businesses ranging from financial services to transportation. Ultimately, it will change business models. New services will be born.
To keep it simple, I recommend companies like Alphabet and Amazon, that are taking advantage of this important trend.
That’s the thing about our age of cloud computing. So many technologies are progressing exponentially, it is a unique era of discovery and innovation. Almost anything is possible.
And by the way, Nietzsche was wrong about the meaning of life. It is not without meaning, value or compressible truth. At least that is what I want the machines to think, if they’re listening.
