AI ‘is not an alien invasion’

I attended in January. He is the world’s pioneer in artificial intelligence and is currently head of engineering at Google.

“This test demonstrated the (AI’s) ability to understand (meaning and) inference, implications, and simple common sense. It cannot beunderestimatedinitssignificance.”

There were no gold medals handed out, but this is truly a milestone and will mean that AI is now going to perform tasks done by most “knowledge workers” in addition to the “blue collar” jobs that automation continues to eliminate.

However, the Stanford University reading and comprehension test showed that the AIs were only slightly better than a native English speaker in comprehension, but the trajectory is obvious.

It provided a glimpse into future capability, said Kurzweil. “AIs can absorb one million books in six seconds and (reading comprehension) will not linger at the human levels, but will improve.”

Both companies were ebullient. Alibaba’s news release credited its software with “topping humans for the first time in one of the world’s most-challenging reading comprehension tests.”

Microsoft said it had made an “AI that can read a document and answer questions about it as well as a person.”

Despite immense capability, the reality is that commercialization, then adoption, of AI technologies usually takes longer than originally estimated.

For example, in 2010, the most high-profile AI product — the driverless car — was supposed to have nearly taken over the world by now. But mass adoption is several years away even though everyone agrees human beings are terrible drivers.

A German report suggested society must first impose necessary operating principles involving driverless vehicles: Vehicles must be hack-proof for public safety reasons and software decisions must place human safety over animals or property if an accident cannot be avoided.

The issue of moral dilemma is more difficult. For instance, does the car swerve to avoid a collision but hit an elderly person or a woman with a baby carriage? The Germans recommend that control of the car must shift to a human in the driver’s seat immediately. If the human fails to act, or is absent, the vehicle must be programmed to stop immediately even if a collision is imminent.

But for those tasks that don’t involve life and death decisions, software is already rapidly replacing paralegals, call centre staff, researchers, tax preparers, journalists writing formulaic financial or sports stories and computer programmers.

On the positive side, social skills will remain in demand or those skills requiring the five senses such as recreational workers, trainers, chefs, psychologists, massage therapists, dentists, physicians, nurses, therapists, social workers, teachers, mechanics, installers, repairers, craftsmen, and designers.

And new jobs will be invented and grow such as app developers, social media managers, driverless car engineers, cloud computer specialists, big data scientists, YouTube content creators, sustainability managers, and drone operators. The fact is that two-thirds of children now entering school in the developed world will end up in jobs that haven’t yet been invented.

“This (AI) is not an alien invasion,” said Kurzweil. “The smartphone is a brain extender. These are tools that enhance our work and personal lives.”

Others have a less generous take. To Alan Blinder, vice-chair from 1994 to 1996, the barons enforced a rigid order that over-filtered information, suppressed internal debate and frequently kept Fed governors in the dark.

“Information is control,” Blinder said. “If some of the governors don’t know about something and the staff does, that’sanadvantageinterms ofmovinginfluencefromthe governors to the staff.”

Under Bernanke, the barons’ grip began to loosen. After Yellen became chair, some measure of internal dissent was even encouraged. Powell’s recent moves are aimed at nudging things further, creating a more relaxed exchange of knowledge.