I can’t remember exactly when it was.
Was it last week when I was at the drive-through ATM?
Was it while I was punching buttons through a phone tree to refill a prescription?
Oh, now I remember – it was while I was at home, selecting a movie to rent online. That’s when it hit me.
Didn’t we used to deal with actual people to get all this stuff done?
I mean, it’s convenient for all of us. When I’m ducking into a supermarket just to get a gallon of milk, I like the self-checkout.
But you see so many folks out of work, and you wonder: Are machines muscling people out of jobs? Or are machines opening up new jobs for people who have to take care of the new machines?
How far can all of this automation go?
Xerox used to have human resource departments for their call centers to help with hiring – to decide whether a person will stay on the job long enough for the company to recoup its training investment. Now, according to The Wall Street Journal, a software program makes those personnel choices.
After years of studying what human academic tutors do to best facilitate learning, companies have modeled computers that mimic successful tutoring techniques.
Remember when the IBM computer called Watson creamed everybody on the TV quiz show Jeopardy? IBM is working with New York City’s Memorial Sloan-Kettering Cancer Center to develop a “decision support tool” based on Watson to help doctors use tons of data to render evidence-based diagnoses that are many times more accurate.
And when you have a technical support problem, you might soon talk to a computer to solve it. A company called IPsoft has spent the past 14 years developing Eliza, a software algorithm that relies on cognitive intelligence to instantly react as quickly as a human brain.
Eliza’s official launch is slated for later this year – and if you’ve ever had to dial customer service to rectify a cable TV problem, you should know that Comcast is one of IPsoft’s biggest customers.
Then there’s this – a team of researchers at the University of Malaga in Spain has developed a computer program that composes contemporary classical music. It doesn’t mimic – the computer cluster called Iamus uses an algorithm that actually creates unique music.
For an average music lover like me – sitting at home playing my kazoo and spinning Spike Jones 78s – this is spooky. So I sought a second opinion from Georgia Southern University professor emeritus of music Michael Braz – a superb pianist and a stirring composer who has taught music literally from here to Kathmandu. Oh, and he played my wedding.
What was the unflappable Dr. Braz’s opinion? He wasn’t very flapped.
“People have been doing that for a while,” he said. “There’s always been some question about the quality of the music, but people have been doing this since the mid-1700s.”
That was when a fascination of math was sweeping Europe, and games emerged that allowed nonmusical people to randomly generate music from pre-composed sections. To choose the musical phrase, you’d roll a pair of dice. The most famous version of the game was attributed to Mozart, though that hasn’t been proved.
I listened to one trio by Iamus, which to me sounded a bit like something modernist composer Charles Ives would’ve scribbled on a cocktail napkin. But as Dr. Braz reminded me, judgment is arbitrary – even when the human touch is missing.
“Anything like this that gets the public continually fascinated with music from another angle is not necessarily a bad thing,” he said.
Of course, I’ve always thought my job was relatively safe from being poached by a computer – until this bit of news scared the pudding out of me: A company called Narrative Science has developed an algorithm that – gulp – duplicates human writing. The operating platform is called Quill, which, according to its website, “can transform data into stories that are indistinguishable from those authored by people.”
You feed Quill financial data or sports statistics and it belches out short, simple stories about business or sports. I read one of its sports briefs, and it creeped me out that I knew it was from a computer – but I couldn’t tell.
A company called ICON Group International takes it a step further with an algorithm that writes entire books. It’s written about a million already, mostly nonfiction and about very specific subjects. But it also does poetry.
And before you ask, yes – I’m actually writing this column.
So which jobs are safe? Ben Goertzel, one of America’s leading researchers in the field of artificial intelligence, told Forbes magazine a couple years ago that the safe jobs are “those that involve transferring knowledge from one area to another, or thinking broadly, creatively and integratively, because these (tasks) require powerful general intelligence, not just narrowly specialized intelligence.”
So. I’ll just need to think creatively with “powerful general intelligence.”