Science Café to Explore AI — and Why Your Computer Doesn’t Understand You (Yet)

November 13, 2017

Machines with morals may soon help people make better decisions, according to Northwestern computer science expert Ken Forbus, who will share his research with the public at this month’s Science Café.

Such technology might help mitigate circumstances like the ongoing Wells Fargo fraud scandal, first revealed in September 2016, in which employees at America’s second-largest bank opened millions of fake accounts without customer consent. “If the institution had software with a stronger moral compass, situations like that might be reduced, because the machine could recognize and question what appeared to be unethical activity,” says Forbus.

But just because robots continue to think more like humans, an uprising by those with Artificial Intelligence (AI) isn’t likely, says Forbus, who holds a doctorate in AI from the Massachusetts Institute of Technology.

Working with other cognitive scientists, Forbus is advancing software that will allow human operators to consider robots as collaborators, rather than merely tools.  “Our interaction with machines should be them adapting to us, rather than the other way around,” he says. “A smart thermostat is a very simple example of this. Another is the autocorrect function on your phone: it rapidly accumulates information about how you use words, including new terms and tries to anticipate what you are writing.”

Technology from the Forbus lab is being incorporated into machines that learn from human language and written sketches. A machine that can adequately assess a sketch could, for instance, compare a student’s answer on a test with a teacher’s solution. Many of these AI advances are based on models of human thought, including analogical reasoning and spatial cognition. The software might also bring moral decision-making to machines, which is important if humans are to trust decisions made by AI.

Forbus’ presentation will also address common AI myths, and discuss self-driving cars and why people may want to postpone widespread implementation of this technology.

“My goal is for people to come away with a more realistic view about what AI is, where it is going, and what the future holds,” he says.

Forbus will discuss his latest research at the November 15 Science Café taking place from 6:30 to 8 p.m. at the Firehouse Grill, 750 Chicago Ave. in Evanston.