In many a popular science fiction film set sometime in the future, humans find themselves at odds with the very technology they developed.
In many ways, that future is already here. We may not be waging war with robots, but the use of artificial intelligence is growing. Are we prepared for all the ways AI will change the way we travel, work, and communicate with others?
Not yet, but we can be, says AI expert Dr. Raj Ramesh.
"I just believe that drawings totally help us understand the world around us and adapt better."
Ramesh is the chief AI officer at DataFoundry, where his work ranges from helping develop code to helping primarily pharmaceutical companies incorporate AI into their business. He also works as a speaker and consultant sharing his insights into artificial intelligence around the globe.
And on Nov. 16 he’ll be at this year’s TEDxNormal event at the Illinois State University Center for Performing Arts to deliver his talk, “How to Think, Transform and Thrive in the Future.”
Ramesh said while humans have been refining computers for decades, the technology has its limits.
Computers do a great job at processing information, but they’re not so good at processing conceptual information—something humans come by naturally.
Ramesh explained computers can only follow the step-by-step instructions laid out by coders. This kind of information processing is linear, much like human language.
“If I read a sentence, I have to have understood the sentence prior to be able to understand this sentence and so on,” he said.
Both humans and computers can generate and process language; but move into the conceptual, non-linear world, and humans have a clear advantage.
“For example, if I recognize your face in a crowd, I can instantaneously recognize that without essentially going through a sequence of steps in my own mind,” Ramesh said. “It’s not like I consciously go through and say well does this person, what's the distance between their eyes, what’s the size of the bridge of their nose, or whatever it is.”
Ramesh explained the step-by-step nature of coding means computers are limited by humans’ understanding.
By feeding them lots of data, he said, we can teach computers to recognize patterns that allow them to identify a face or a chair. But recognizing that the face belongs to your friend, or that the chair would be a good place to take a seat—humans haven’t figured out how to codify those processes.
Ramesh said while it’s theoretically possible, “We have to have more than a couple of breakthroughs to get to the point where we truly can make computers better in some sense cognitively than humans.”
That point is known as the singularity, although Ramesh argued it likely won’t be a single point at all.
There are different aspects of human intelligence that can be surpassed, but not all. And so that’s an open question still in the community, he said.
Until that happens—if it happens—humans need to focus on building their unique strengths to complement the strengths of computers, Ramesh said. Especially if we want to stay employed.
“Most of us are concerned in some way or the other, either for ourselves or our kids or for the next generation, that computers might take over our jobs,” he explained. Already machine automation has changed the face of manufacturing, customer service and even healthcare.
But because humans’ strengths differ from those of computers, “There are pieces of the job that will not be taken over by computers, or there will be new jobs created which will require a higher level of thinking and skills, that we should seek,” Ramesh said. “We humans should try to fill in that gap, and not just be afraid of what AI or computers can do.”
Ramesh has a simple suggestion for how to get started: just doodle.
He explained visual representations are easily understood by humans. “We are visual learners,” he said. And when we use visual models to communicate, we don’t need to rely on linear logic to make and understand meaning.
“By virtue of drawing something on paper, I’m taking my mental model and translating it on paper, and amazingly this same mental model seems to be going into the mind of the listener and observer,” Ramesh said.
You don’t have to be an artist to make doodling work for you. “When kids go to school, from kindergarten, we have a natural ability to draw, and that’s how we express our thoughts,” Ramesh said. But by focusing on linear processes like written and oral language, schools train students out of using this tool, he said.
Still, it’s never too late to get started, Ramesh said; in fact, Ramesh uses simple visual representations to teach complex ideas in his YouTube videos. He said he’s been working on the concept of doodling for better understanding for over 10 years, and thinks anyone can make the tool work for them.
“You should try it yourself as well, no matter what your area of expertise is. Mine just happens to be artificial intelligence. It could be in medicine, it could be in plumbing, it could be in any wide variety of topics. I just believe that drawings totally help us understand the world around us and adapt better.”
Tickets for TedxNormal 2019 are $20 or $10 with student ID and are available online or in person at the CPA box office.
People like you value experienced, knowledgeable and award-winning journalism that covers meaningful stories in Bloomington-Normal. To support more stories and interviews like this one, please consider making a contribution.
