IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Humans Need to Teach AI for Best Results

With some two-thirds of businesses and IT organizations eyeing artificial intelligence deployments in the coming year, making sure they perform will mean lots of training.

(TNS) — Everyone knows about child rearing, but the big career of the future may be computer rearing.

Machines with artificial intelligence cannot teach themselves, even with the advent of neural networks and machine learning. They will still need tutors to steer them to accomplish a given task efficiently.

The growing human-machine symbiosis — think smartphones — will change every aspect of our lives and how our children make their living. The time to prepare is now.

Two-thirds of business and information technology executives plan to invest in artificial intelligence, or AI, this year, according to a global survey of 6,300 senior managers by consulting firm Accenture. About 75 percent of Houston executives said they would invest in AI next year, and they expect machines to operate alongside humans at their organization as a co-worker, collaborator or trusted adviser.

“We want to augment folks so that they are focusing on the highest-value task, that they are taking the optimum route, and they have the most information and insight that is possible,” Brian Richards, head of Accenture’s Innovation Hub in Houston, told me.

“As a business, your ability to consume AI, and train your workforce to consume it and deploy it and use it to improve your competitive position isn’t something that’s going to last only the next few years. It’s going to go on for decades.”

Every day a new machine hits the market that can replace a human employee, spreading anxiety across the workforce. And those concerns are valid. The McKinsey Global Institute, the consulting firm's think tank, found that up to 800 million people, including a third of Americans, could lose their jobs to machines by 2030.

While many of those jobs are blue-collar, such as truck driver, white-collar professionals such as lawyers and bankers will see the rote aspects of their jobs performed by machines that can better analyze data and generate recommendations.

Artificial intelligence is already in our lives. Machines decide the best way to navigate a city during rush hour using the latest traffic data. Digital assistants on kitchen countertops allow consumers to order exotic items from every corner of the world with a voice command.

In the workplace, AI is intended to augment human employees. It will allow a company to sell more products and services, employ fewer workers or do both.

AI, though, will need to be trained and managed, like any other employee, creating new jobs and new opportunities, said Mike Redding, managing director of Accenture Ventures, the consulting firm’s investing arm.

“We have to make sure it is making good decisions, appropriate decisions. We’re suggesting that you should think of it as a child, and to have it play an appropriate role, the organization has to be the parent,” Redding said.

“There is good parenting and bad parenting. You really need to think about the concept of raising it to know between right and wrong, what to do when it encounters structured situations, and what to do the first time it encounters something it’s never seen before.”

Computers can teach themselves to identify patterns, such as learning to read traffic signs, and then act accordingly. But context is important. Redding described an AI that analyzed millions of photos and ultimately concluded that every person who stands next to a stove is a woman because that’s the pattern it discovered.

Companies need to teach and test their machines to make sure they don’t draw the wrong conclusions.

“Now you have to have a learning division for your bots. It’s a new job: bot trainer,” Redding said.

AI could also help displaced workers transition to new careers. A worker wearing a headset with an earphone and an augmented reality display will be able to access expert knowledge instantly.

“The net result is that we can take someone who has done one thing their whole life, and we can pivot it and take those core capabilities and reapply them,” Redding said. “Through augmentation, we can make people more adaptable.”

Companies will also need more cybersecurity personnel because hackers will not only steal data but replace it with harmful data. More than 80 percent of Houston executives agree that business organizations are not prepared for AI using bad data and behaving badly.

The real question is whether our schools and culture will produce a workforce prepared to leverage AI and capture the potential boost in productivity and economic activity. The key to a bright future is to learn how to work with machines, not against them.

©2018 the Houston Chronicle Distributed by Tribune Content Agency, LLC.