How Augmented Decision Making is Changing Everything.
In previous Insights posts, I’ve described the promise and potential downside of future artificial intelligence (AI) breakthroughs. Stephen Hawking, Bill Gates and Elon Musk are among tech and science luminaries who have expressed grave concerns about the possibility that by mid-century, AI that could spin out of control, quickly surpassing human capabilities.
Since its modest beginnings in the 1950s, we’ve experienced cycles of optimism and pessimism about AI’s potential. What most casual observers miss entirely is the parallel, but distinct evolution of research and development in two areas—intelligence augmentation, which works to support our decision-making abilities, and artificial intelligence, which works to develop autonomous digital decision-making. Significantly, intelligence augmentation has received much more funding and achieved more consistent beneficial results than AI.
There is a clear distinction between the two–
“Artificial Intelligence” attempts to replicate human cognition in a computer system to advance such systems towards human-like autonomy, i.e., to create digital entities capable of taking independent action in different environments.
“Intelligence Augmentation,” by contrast, is based on the proven premise that we can enhance computer systems to assist human thinking, planning and other forms of decision-making—leaving humans in complete control of the objectives to be achieved. In other words, intelligence augmentation relies on human-computer interaction (often referred to as HCI), rather than autonomous digital systems.
AI holds the potential for one day creating things like completely autonomous cyber soldiers and drones. HCI, on the other hand, is already working to develop exoskeletons that help paraplegics walk and increase human strength for able-bodied workers. Its greatest future benefit, however, will be helping us make making better-informed, quicker decisions.
Can the Human Brain Be Digitally Replicated?
The underlying assumption of AI is that intelligence can be mathematically modeled with highly sophisticated logarithms that will in the future be able to replicate the subtle nuances of human thinking. To date, however, this remains an elusive goal for scientists who work at the nexus mathematical modeling and computer systems.
Unfortunately for AI enthusiasts, there is yet no evidence that the human brain can be simulated mathematically—especially in areas like human intuition, processing ambiguity and situational contexts. Adding to the challenge, human intelligence also encompasses non-rational bodily and perceptual processes.
Google Is Based on HCI
Researchers hope to develop AI capable of ‘zeroing in’ on a problem quickly without needing to process on all contingencies. It’s not clear, however, how computers could ever do this.
By contrast, HCI-supported “counting out’ abilities, e.g., finding the best answer among a vast set of possibilities from cloud-based big data—far exceed human capabilities. HCI, of course, is the basis for Google and all other search software.
In Part II of this article, I’ll describe how HCI is going to accelerate our ability to make all kinds of decisions with much improved outcomes.