White House Chronicle

News Analysis With a Sense of Humor

My Poetic Quest to Understand Artificial Intelligence

by Leave a Comment

Myself when young did eagerly frequent
Doctor and Saint, and heard great Argument
About it and about; but evermore
Came out by the same Door as in I went.

I feel close to Omar Khayyam, the great 11th-century Persian poet and mathematician, not just because of his fondness for a drink but also because of his search for meaning, which took him in “The Rubaiyat” to “Doctor and Saint” and then out “by the same Door as in I went.”

I’ve been looking at artificial intelligence (AI) and I feel, like Omar, that I’m coming away from talking with leaders in the field as unenlightened as when I started this quest.

The question is simple: What will it do to us, our jobs and our freedom?

The answer isn’t clear: Even those who are enthusiastic about the progress they’re making with AI are privately alarmed about its consequences. And they worry about how far some corporations will push it too hard and too fast.

The first stages are already active, although surreptitiously. The financial technology (fintech) world has been quick to embrace AI. Up for a bank loan? Chances are you’ll be approved or turned down by a form of AI that checked your employment, credit score and some other criteria (unknown to you) and weighed your ability to repay. Some anomaly, maybe a police report, may have come into play. You’ll be told the ostensible reason for your rejection, if that’s the case, but you may never know it.

The two overriding concerns: what AI will do to our jobs and our privacy.

If jobs are the problem, governments can help by insisting that some work must be done by human beings: reserved occupations. Not a pretty concept but a possible one.

When it comes to privacy, governments are likely to be the problem. With surreptitious bio-identification surveillance, the government could know every move you make — your friends, your business associates, your lovers, your comings and goings — and then make judgments about your fitness for everything from work to liberty. No sin shall go unrecorded, as it were.

This one isn’t just a future worry, it’s nearly here. The Chinese, I’m told, have run an experiment on citizen fitness using AI.

Historically, at least in literature, we’ve been acculturated to the idea of man-made monsters out of control, whether it was Mary Shelley’s “Frankenstein” or Robert Louis Stevenson’s “Dr. Jekyll and Mr. Hyde.” But the mythology probably has been around since man thought he could control life.

On jobs, the future is unclear. Until this point in time, automation has added jobs. British weaver Ned Ludd and his followers, who smashed up the looms of the Industrial Revolution, got it wrong. Nowadays cars are largely made by machines, as are many other things, and we have near full employment. Fields like health care have expanded, while adding technology at a fast pace. AI opens new vistas for treatment.

Notoriously difficult-to-diagnose diseases, like Myalgic Encephalomyelitis, might be easily identified and therapies suggested.

But think of a farm being run by AI. It knows how to run the tractor and plow, plant and harvest. It can assay the acidity of the soil and apply a corrective. If it can do all that, and maybe even decide what crops will sell each year, what will it do to other employment?

In the future AI will be taught sensitivity, even compassion, with the result that in many circumstances, like customer assistance, we may have no idea whether we’re dealing with a human or AI aping one of us. It could duplicate much human endeavor, except joining the unemployment line.

I’ve visited MIT, Harvard and Brown, and I’ve just attended a conference at NASA, where I heard some of the leading AI developers and critics talk about their expectations or fears. A few are borne along by enthusiasm, some are scared, and some don’t know, but most feel — as I do, after my AI tour — that the disruption AI will bring will be extreme. Not all at once, but over time.

Like Omar, I came away not knowing much more than when I began my quest. “The Rubaiyat” (which means quatrains) is a paean to drink. At least no one suggested machines will be taking to the bottle, but I may.

You Need to Be Brave for This New World

by Leave a Comment

Pondering the future requires an extrapolation from a data point in the present. But different data points give very different futures. Beware of the prognosticators.

Take this as a data point: Stephen Entin, senior fellow at the Tax Foundation, a think tank devoted to tax studies since 1937, predicts that with an aging population and low birthrates, we’re going to need more immigrants to fill the federal and state coffers with their taxes. We’re also going to need hundreds of thousands of workers for health care and aged care in the years ahead, he says.

Or take this as a data point: MIT Sloan Professor Tom Kochan fears that artificial intelligence will substitute for millions of employees. Retraining is possible, but can you see a long-haul truck driver pushing wheelchairs in an assisted-living facility? Not easily.

Upheaval in work is the most predictable aspect of the future.

It is, if you will, already arriving in the workplace. New techniques and new concepts of what is work are afoot.

The old concept is that a person leaves school, gets a job and signs on to the social/work contract — gets company-paid benefits and expects security and stability. The infrastructure of society pointed the way to employer-employee model.

The new concept is the gig economy, where contract work and freelancing rule. The work/social infrastructure where medical insurance, Social Security and retirement are part of the deal is dying. But a one has yet to emerge in concept and in law.

Business is in the throes of its own future adjustment. Take 3D printing, more correctly called additive manufacturing. What was novelty a decade ago is now a tool used in industrial plants across the country. Instead of taking a chunk of metal, say aluminum, and cutting and lathing it to make a part, which wasted most of the metal, there’s no waste with 3D printing.

Now to make a part, you print it from metal powder to a design lodged in a computer. The saving in material, shipping and manpower is enormous.

And additive manufacturing, just like everything else on the shop floor, can be automated. Machines can sinter — the term for 3D printing — through the night with only artificial intelligence supervision.

There’s a new existential worry in every large enterprise in the United States, from banking to manufacturing, from electricity generation to hospital management and from building crane operation to pharmaceutical design: cyber-vulnerability.

To paraphrase Leon Trotsky, you may not be interested in cyber-war, but cyber-war is interested in you.

I’ve interviewed widely on the subject, from top academics to some of the most successful cyber-security entrepreneurs, to National Security Agency sources. The story is the same everywhere: Nothing connected to computers is entirely safe; and if it’s safe today, will it be tomorrow? That plague, like the plagues of old, will, I’m assured, be with us for decades, if not centuries to come.

Cyber-defenders build, cyber-hackers build around. It’s a version of what one secretary of defense, Harold Brown, said about the Soviet threat in the Cold War: “We build, they build.”

The changes are all around the home: Everything has changed since the day of the black AT&T phone, but you haven’t seen anything yet. Your packages may be delivered by drone, your phone service will be entirely mobile, and your life will be dictated by electronic secretarial aids. Alexa is just the beginning. With artificial intelligence, these robots will talk back to us and maybe argue, shudder the thought.

I pity the dogs. We had a dog that would be very upset if she heard my wife, a talk show regular, on the television when she was also elsewhere in the house. Dogs are sensitive to these things.

What if man’s best friend, eternal unquestioning companion, develops a strong affection for the electronic assistant and changes loyalties, especially if the gadget is feeding the dog? Will it be as Julius Caesar might have said, “Et tu, Fido?”

 

 

Photo: Photo by rawpixel on Unsplash

Skilled Jobs Go Begging Now, But Thinking Machines Are Coming

by Leave a Comment

Consider it as the work dichotomy.

There is a shortage in the millions for skilled labor jobs in the United States. The country is desperate for men and women who drive trucks, operate machines, weld, wield hammers — or can fill skilled jobs in dozens of categories from bulldozer operator to utility lineman.

Bill Hillman, chief executive officer of the National Utility Contractors Association, the organization that represents contractors (people who do everything, from replacing electricity poles to working down manholes to operating heavy equipment), says getting help is a major problem for his members. So they are setting up training programs and working with schools and community colleges.

But these also are some of the people who could be jobless due to artificial intelligence (AI) in the near future. Thomas Kochan, co-director of the MIT Sloan Institute for Work and Employment Research, told me this “middle of the labor market” is coming under attack by AI deployment.

John Savage, professor of computer science at Brown University, foresees a need for major retraining of workers with the spread of AI. But he told me he is “optimistic”: He sees major displacements but new opportunities.

Displacement is a worry for workers, but so is job quality deterioration in the so-named gig economy or freelance economy: a volatile labor pool where the employer holds most of the cards.

Gig workers are spread among diverse occupation groups: arts and design, computer and information technology, media and communication, transportation and material moving, construction and extraction. They are working here and there without permanence, medical insurance or pension provisions, like employer 401(k) contributions.

That is for starters and it is happening now. Then comes the apocalypse when millions of workers find themselves displaced by thinking machines. Think of what happened to elevator operators in cities when elevators were automated.

The first to go might be taxi drivers, some truck drivers, airline pilots and others in transportation. Already in Phoenix, you can ride in a robot taxi operated by Waymo, the Google self-driving car project. Truck makers, stirred on by potential competition from new entrants, like Tesla, are hard at perfecting autonomous intercity trucks.

To my mind, the issue is not whether but when. There are more than 3 million truck drivers on U.S. roads. Not all will be displaced by AI, but if 1 million go, there will be considerable downward pressure on wages.

Traditionally, and Savage points this out, automation has led to a surge in new, different jobs. Ned Ludd, who with his followers destroyed mechanical weaving machines in England in the early 1800s, was wrong. Mechanized weaving added far more related jobs than those lost.

But this time it could be different, warns John Raymont, chief strategy officer of Kurion, an advanced technology nuclear company. He says the difference is that automation heretofore has led to more products, and therefore more jobs. Artificial intelligence threatens to take away jobs without producing new products, which themselves produced new jobs.

Take the automobile production line: It led to more people being able to afford cars and more jobs maintaining and fueling those cars. It enhanced America’s growing prosperity.

So far, AI appears to be aimed directly at employment. In the way that cheap labor in Asia sucked manufacturing jobs out of the United States, so machines may take over skilled jobs from airline pilots to Uber drivers, Raymont says. Other jobs may still be safe, including plumbers, he says.

And it will not be just manual workers who will have their jobs taken over by wily computers. Accounting, tax preparing and auditing, money lending, loading and unloading ships and trucks will be done by machines guided by artificial intelligence. A ship, it is theorized, will be able to leave a U.S. port without the aid of seamen or dock workers and sail to Singapore, dock and unload autonomously.

Job displacement may have this opportunity: More leisure time in which people can play golf on greens maintained by thinking mowers, aerifiers and fertilizer spreaders. After they play, a machine may make them an extra dry martini at the club bar.

 

Photo: August 6, 2017 Mountain View/Ca/USA – Waymo self driving car cruising on a street, Silicon Valley. Editorial credit: Sundry Photography / Shutterstock.com