Andrew McAfee: it’s amazing how bad machines are at common sense

Lecture in Abu Dhabi hears from tech and artificial intelligence expert on machines, AI, jobs and the future

ABU DHABI, UNITED ARAB EMIRATES -January 31, 2018: Andrew McAfee, Principal Research Scientist at MIT Sloan School of Management (C), delivers a lecture titled "The Future of Jobs in the Age of Artificial Intelligence", at Majlis Mohamed bin Zayed.

( Rashed Al Mansoori / Crown Prince Court - Abu Dhabi )
Powered by automated translation

In March 2016, a Google computer program took on the world’s greatest player at the ancient game of Go. Lee Sedol had an unpredictable, brilliant style – the hardest person for a computer to beat.

But in a game that had echoes of chess master Garry Kasparov’s battle against Deep Blue, the machine won 4-1.

“I kind of felt powerless,” Lee said, after going 3-0 down.

Depending on which side of the fence you sit, this is either progressive or frightening.

This example was part of an engaging talk in Abu Dhabi yesterday by Andrew McAfee, a technology and artificial intelligence expert at Massachusetts Institute of Technology.

“The future of jobs in the age of artificial intelligence” touched on many topics such as robots, jobs and inequality.

Dr McAfee began with a sobering assessment of how computers are making rapid advances in the intelligence stakes.

“Are machines doing useful work? Yes,” he said. Microsoft now have a machine that can understand speech better than human beings can. Computers can diagnose diseases better than us, while Google uses AI to cut energy bills at data centres by as much as 13 per cent.

“What do we need human minds for,” he asks. The good news is, a lot. “We’re good at things the machines are terrible at – it’s amazing how bad machines are at common sense.”

He points to the scandal in Sydney in 2014 when Uber’s fares surged after a terrorist attack in the city.

“It was the wrong thing to do and Uber made an apology. No human would have made that choice, but its algorithms did .”

A second point is the human ability to ask questions and here, Dr McAfee gave another example. As Pablo Picasso said of computers: They are useless. They can only give you answers.

“Even more valuable is our ability to ask a question,” Dr McAfee said. “What is the next problem to tackle? How do we make progress? This is important as asking questions is going to become more valuable.”

A third is that computers still have poor social skills and finally, Dr McAfee believes people and computers are learning to work together. He points to the Shanghai Tower in China, which was first drawn up by a computer and then improved by humans. “We improved it and do what we do best.”

Dr McAfee is a research scientist at MIT, and studies how computers are transforming the world around us. His 2014 book Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, co-authored with Erik Brynjolfsson, was a top 10 bestseller.

Dr McAfee’s talk was well suited to the UAE. Last year, in a world first, Omar Al Olama was named as Minister for Artificial Intelligence. The UAE’s strategy on AI was unveiled in October and Dubai wants to have one in four journeys made by autonomous vehicles by 2030.

But, more broadly, how can we all prepare for this brave new world of the machines? Fortunately there is quite a lot.


Read more:

Year in review: UAE embraces artificial intelligence 

World's first self-flying taxi coming to Dubai 'within months'


According to Dr McAfee, humans and machines can team up to explore new territory. Here he suggests having a human doctor train with a machine and learn from the diagnoses that they make. “Having humans diagnose the disease first will seem really strange and crazy in years from now.” Dr McAfee gave  an example of a job that will exist in the future – a health coach. The person will check if you are following medical advice, while the machine will diagnose the problem.

Machines and computers can also improve human performance. After Mr Lee’s defeat, he now wins 95 per cent of his games – up from 75 per cent.

“They will make us smarter, not dumber.”

But an area of danger is the future of warfare and the possibility that humans could built lethal autonomous weapons.

“It then becomes scary. We have to face these issues.

“I’m an optimist but we have to be careful.”