Menu
Menu
Now that AI has mastered 'Go', are all our jobs next?

Now that AI has mastered 'Go', are all our jobs next?

'There is disruption coming,' admits the CEO of one AI company

There was once a time when no one thought computers could master chess; then, in 1997, IBM's Deep Blue beat chess champion Garry Kasparov. The bar then moved to the ancient Chinese game of Go -- until Europe's reigning human champion fell to Google's AlphaGo system late last year.

One by one, artificial intelligence has overcome the obstacles set before it. Is this all part of an inevitable trend leading to humanity's obsolescence -- or, at least, unemployment?

"Absolutely, there is disruption coming," said Shashi Upadhyay, CEO of Lattice Engines, which uses AI for predictive marketing.

Just last month, a World Economic Forum report warned that AI, robots and other tech advances will take more than five million jobs from humans over the next five years.

Here's an example: There are currently more than 230,000 taxi drivers in the U.S., according to the Bureau of Labor Statistics -- not counting Uber or other alternatives.

"Once driverless cars are mainstream, all those jobs are going to go away -- those people are going to have to find something else to do to support their families," Upadhyay pointed out. "That's disruptive."

It doesn't mean "the machines" are on a quest to destroy our lives, however.

"I don't buy into the idea that machines have purpose and a value system and are somehow out to destroy the world," said Upadhyay, who was formerly a data scientist at Cornell University. "I think we'll coexist."

In fact, AI can in many cases free humans from tasks they're not particularly well-suited for in the first place and allow them to concentrate on what they do best, he suggested.

Marketing -- the focus of Lattice Engines -- is one example.

"When we first introduced our predictive-intelligence product in 2011, it could probably outperform about half the people in a sales organization," Upadhyay said.

Specifically, the system was better than roughly half at predicting when a prospect would make a purchase.

Five years later, the technology has "seen" so many more examples of who buys and who doesn't that it can now outperform roughly 90 percent of salespeople at that prediction, he said.

"In a way, it's looking through data from 20 million U.S. businesses -- that's something no sales rep can ever put in their head," Upadhyay said.

The effect, though, isn't that companies get rid of all their salespeople and replace them with predictive-marketing software. Rather, it's changing the human focus from predicting who will buy to closing the deal, he pointed out.

"It's moving from one skill that people were bad at anyway and shifting to what they're good at," Upadhyay explained. After all, marketers tend to be creative -- "in cases like this, the machine is actually freeing people to do what they love."

That's a pattern Upadhyay expects to see repeated.

"My belief is that machines will do some things better, and we'll continue to do other things much better," he said.

Humans tend to have very poor intuition where small numbers are involved, for example, so that's an area in which technology will likely reign supreme.

"If I tell you that there's a 75 percent chance it's going to rain versus a 25 percent chance, you know what that means," he explained. "You'll bring an umbrella."

Forced to compare a 0.1 percent chance someone will buy a product versus a 0.4 percent chance, however, humans have a hard time -- machines can do much better.

Advertising conversion rates are a perfect example, he said.

Machines also tend to excel any time having lots data is an advantage, but -- conversely -- humans shine when sample sizes are small.

"Say you're a rep with only 20 accounts, but you need a lot of detail about those people," Upadhyay explained. "There's no question humans will outperform machines in those cases."

Part of the reason for that is that human interactions tend to be based around stories, and while machines are great with lots of data points, "we still don't have machines that understand stories," he added.

Accordingly, things that have usually been the domain of the humanities and social studies will remain dominated by people, he predicted.

Ultimately, any technology is essentially amoral, so it's all a matter of how people use it, Upadhyay said. What's important is that those uses are thoughtfully discussed.

The use of robots in wars is one that needs particularly close attention, he added.

But with the likes of Stephen Hawking and Elon Musk already weighing in on AI, "I'm optimistic, because this topic has the attention of all the right people," Upadhyay said.

Still feeling anxious about it all? Upadhyay recommends learning more.

"Most of what's called machine learning today works on ideas that you can explain to a 10-year-old," he said.

Skills in basic programming and statistics can make a lot of it less mysterious and less scary, he added.

"Just knowing a bit about how these things work will go a long way toward helping you understand that this trend is not something to be scared of," Upadhyay said. "Humanity has a lot of things to be proud of -- being 'Go' champion doesn't need to be one of them."

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO New Zealand newsletter!

Error: Please check your email address.

Tags predictive analyticsLattice Enginesdata-driven marketingemerging technologyartificial intelligence

More about GoogleUber

Show Comments