For many of us, the concept of artificial intelligence conjures up visions of a machine-dominated world, where humans are servants to the devices they created. That’s a frightening image, inspired more by Hollywood and science fiction writers than technologists and the academic community. The truth is less sensational but far more meaningful.
We’re actually nowhere near the self-sustaining robots Isaac Asimov imagined in I, Robot. What we have instead is intelligence amplification (IA), a field with exponentially more potential to change the world in the immediate future.
The distinction between AI and IA is as simple as it is significant. AI makes machines autonomous and detached from humans; IA, in on the other hand, puts humans in control and leverages computing power to amplify our capabilities.
For a real-world example of IA, look no further than IBM’s Watson, an intelligence amplification machine that is often mistaken for AI. The feedback loop created by exposing intelligence to humans through APIs enables Watson machine to learn and improve the information it provides. The machine presents that information to humans and then learns from their decisions. Like much of IA, Watson becomes smarter by amplifying our own intelligence.
While humans have used tools to bolster their productivity for centuries, the proliferation of application programming interfaces (APIs)—the mortar connecting the bricks of our digital world—in recent years has enabled greater access to valuable information in real time. The combination of intelligent computers, intelligent software, and APIs has profound implications for our everyday lives.
Doctors, for example, stand to benefit tremendously from IA in their interactions with patients. Say you have a doctor at the Mayo Clinic making a diagnosis. The patient is relying on the doctor’s expertise—but the publication of new medical research far outpaces the doctor’s ability to consume and analyze it. That’s where IA comes in. Rather than depending on his or her finite body of knowledge, the doctor can utilize supercomputers capable of surveying vast amounts of information quickly to present decisions the doctor might not have thought of or known about.
Meanwhile, present-day robots can hardly stay upright.
This isn’t to say artificial intelligence doesn’t have a significant role to play in the evolution of intelligent computers and they way we interact with them. Researchers at MIT, the University of Toronto, and elsewhere have advanced AI’s value in performing “soft intelligence” tasks like facial identification and pattern recognition—activities that ultimately improve judgment across the entire system. However, when it comes to “hard intelligence” activities like driving a car, AI still has a lot of learning to do.
Visions of the future have distracted us from what’s possible today. While Google experiments with self-driving cars that can be derailed with a simple laser pointer, automakers around the globe have already begun introducing IA-enhanced cars that can improve safety by assisting drivers with duties like highway driving on long-distance road trips. Tesla, Volvo, and Audi have or will soon introduce “autopilot” functionality on their vehicles. Though it’s still unclear when autonomous vehicles will become affordable for most Americans — keeping them in a world of moonshots for now — IA-integrated cars are something we can advance, utilize, and benefit from today.
Of course, technology will always need moonshot ideas —they're what makes humans great. But focusing too heavily on fully-formed artificial intelligence misses the great strides we’re making here and now with intelligence amplification that’s actually changing lives.
The future of machine collaboration we’ve fantasized about is already here, and it's not what we've been taught to fear. Our machines really are here to serve us—all we have to do is embrace them.
