Friday, 20 November 2015
SMARTS Update: When They Praise the `Smart` Future, Follow the Money
Last Sunday morning, I was puttering around in the kitchen with the radio droning in the background. Jim Brown of CBC`s The 180, produced out of Calgary, began talking to a British man, Daniel Susskind, who is promoting a book he has co-written with Richard Susskind (his father according to The Times of Israel) about the reinvention of the professions by smart machines. It's called The Future of the Professions, published by Oxford University Press, priced at £18.99.
Brown opened the segment by quoting an unnamed US study which says 45% of all the jobs humans do now could be done as well or better by machines using current technology.
Anything to do with machine intelligence catches my attention. My new book SMARTS traces out how, for millennia, we described ourselves as the only intelligent beings on the planet; then realized that all of nature is intelligent; and now mimic all sorts of different forms of intelligence in all kinds of smart machines. The two main thinkers behind this SMARTS revolution are Charles Darwin and Alan Turing. More on Turing later.
Brown described Susskind as a Lecturer at Balliol College, Oxford, one of the world's finest educational institutions. When academics are introduced in the media or in law courts as experts, we assume that their positions mean that their work and their opinions are independent of corporate interests, especially when their employer is as prestigious as Oxford. Oddly, Susskind's specific expertise was not mentioned. (It's economics.) Brown plunged instead to the book`s substance. Apparently the Susskinds argue that most of the professions will be utterly transformed by smart machines and algorithms and we should not fear this, no, we should welcome it. Why? The professions exist to solve problems, not the other way around. While smart machines that learn on their own will solve these problems differently than humans, they will certainly do it at least as well (and definitely cheaper).
Brown said but hey, won't that mean that radio interviewers and even university Lecturers may soon be out on the street? Susskind responded with a certain upper class tone that silently screamed ' you don`t even know what questions to ask,' though his words went in another direction. This won't happen so fast, he assured Brown. At first, lawyers and doctors will just do high level things as machines take over the grunt work. Human professionals will remake themselves as " knowledge engineers" whose expertise will be mined as the machines get up to speed. This has already been done with the accounting software that most people use nowadays to file taxes.
Neither Brown nor Susskind mentioned that machines that learn will not need knowledge engineers for long, they'll become their own knowledge engineers in short order. This is when I started to yell at the radio. Ask him about that, Jim, I said. But Jim couldn't hear me.
The real issue, Susskind continued, is that doctors and lawyers and accountants should not be motivated by a desire to earn their livings solving people's problems, but rather by the desire to solve the problems.
Should not is not the same as will not, I yelled, thinking of my Dad who threw himself into the famous Medicare strike in Saskatchewan to maintain doctors' rights to earn fees for services and hold onto their independence.
If there are better ways to solve problems involving robots and algorithms, why not use them, continued Susskind.
Brown did not respond with hey wait, sometimes the way you solve a problem creates new ones that are worse, like the way we solved the problem of Saddam Hussein. Instead he moved to the next issue: what of the emotional support professionals give clients? Wouldn't you want a human doctor to tell you you're dying instead of a machine? Won`t machines fail at empathy?
Susskind dismissed this concern on the grounds that machines can identify real human pain better than humans can. Then he turned his own argument around 180 degrees and asserted that the choice is rarely between the best and the lesser, but between the lesser and nothing at all. As an example, Susskind pointed to the billions of dollars invested by Japanese conglomerates to build smart robots to take care of its huge population of the infirm elderly. Susskind finished off with a quote from Voltaire that Perfection is the enemy of the Good. [According to Wikipedia, Voltaire was quoting an Italian proverb]. In other words, it's better to get the job done badly then not to care for these old people at all.
Hah, I said to the radio. Another famous thinker, Bertrand Russell, showed that from any false premise he could prove a false conclusion. (It's the Pope and I are one proof, see it here). The false premise in this argument is that there are no alternative caregivers to be found. But the Japanese wouldn't have a caregiver problem if they allowed foreigners from poorer countries to immigrate, people who would be happy to do these jobs. The unnamed problem here is that the Japanese dislike of foreigners is so profound, machines are preferred.
Now why wouldn't he mention that? I asked my husband. He ignored me. He was buried in the New York Times' coverage of the terrorist killings in Paris.
That's when I realized that Susskind had not answered Brown's first question: what has happened, or will happen to those pushed from jobs by the onslaught of the smart future? What will result when the young professionals who owe a fortune in student loans have to line up at the Food Bank?
I was still stewing on this the next day when " the economy of plenty" was mentioned in the comments section of my previous blog post. A technical writer named Ben Williams wrote that my post had implied it. I thought this idea had vanished with the long defunct Canadian Social Credit Party which used to argue that the Canadian economy would be just dandy if the government printed more money.
What are you talking about, I wrote back.
His response boiled down to this: in the coming age of smart machines, it will not matter that machines write stories and make art. This will not prevent any humans from making these things too. There will just be more interesting works out there to enjoy and that will be terrific.
A nice premise, unfortunately false. Most economists will be happy to explain that scarcity of something in demand will drive up its value, while ubiquity will reduce it. If a commodity isn't actually scarce, some people will try and make it so. (The Hunt brothers famously tried and failed to corner the market on silver which for a time drove up its price beyond all reason.) The same holds true for services. This is why self-regulating professions try to control the number of entrants to their fields. Journalists are not professionals but we used to be both scarce and necessary, and so we could earn a reasonable wage as our publishers made money by selling advertising and subscriptions. With the rise of free platforms, free distribution systems like the Internet and Wi-fi enabled phones that upload and download text and video, and the consequent rise of citizen journalists, the economic value of professional journalism began to collapse. This is what lawyers and doctors can expect when smart machines become available at low cost to perform the services they provide for significant fees. Just last week, IBM took out an eight page ad in the New York Times to announce the dawn of this new 'cognitive' age and the availability of IBM machines with the software to manipulate Big Data which will out-think humans in almost every area of endeavor.
The people who will make money from this economy of plenty will be giant international firms that can afford to get rid of human staff and buy machines instead. Those who will gain the most are the owners of the corporations bringing smart machines to market first, companies like IBM, Google, Apple, Microsoft. Historically, the bigger the company, the harder it is to keep innovating, which is why Google, Apple, etc. have been buying novelty at a furious pace (in the form of startup companies with the good robotics patents and professors with clever artificial intelligence algorithms). Google has an endless supply of cash with which to make such purchases. According to its last annual report, it sits like a dragon on a pile totaling more than $60 billion.
They are trying to buy control of the future, to corner the market on artificial intelligence and the smart robots they plan to sell to us. Selling involves experts telling you how great something is. Selling involves the publication of arguments in favor of one particular future, one that has to be made to appear inevitable as well as desirable.
Hey, I said to my husband: you don't suppose that either Susskind is connected to companies with interests in this stuff?
Jim Brown had not asked that question on air.
And so I got down to it. First I went to Richard Susskind's home page. It carries bios of both authors as well as information about their book. These two are, to put it mildly, connected. Daniel has two degrees in economics from Balliol College, Oxford. He has worked in the Prime Minister's Strategy Unit and in the Policy Unit of 10 Downing Street, as well as for the Cabinet office. Richard, the father, has a deep and abiding interest in machining the law. He too is an Oxford product, with a doctorate in law and computers from Balliol College. He has been an IT adviser to the Lord Chief Justice of England and Wales, written many articles and books, is an Honorary Bencher of Gray's Inn, a Fellow of the Computer Society, the Royal Society of Edinburgh, etc. And oh yes, he serves as Chairman of the Advisory Board of the Oxford Internet Institute (more on OII below.)
His bio also records that " although Richard is self-employed and works independently, he does not claim to be a dispassionate analyst or to be free of commercial interests."
Ah, I say to self.
The next page reveals that Richard works in the commercial as well as the academic world, that he consults with major law firms, "and to in-house legal departments," that he has been "an active member of an advisory board of Lyceum Capital, a private equity firm that is committed to investing in the legal profession; and he was chairman of the advisory board of Integreon, a legal and business process outsourcing business." Richard apparently discloses conflicts whenever they arise.
Integreon is part of a huge conglomerate (which controls 20% of the value of the Manila stock exchange) owned by the Ayala family who have helped run the Philippines, with its extreme divisions between rich and poor, since the 19th Century. Integreon offers machined office processes, including legal processes, world wide. Lyceum makes no mention of Susskind but I take his word for it that he has advised them.
Oxford itself is knee deep in funding from a major party involved in the development of artificial intelligence products -- Google. In 2014, Google purchased Deep Mind, a UK startup which had published some startling artificial intelligence results in Nature, for $400 million. Deep Mind`s work is similar to that of Chris Eliasmith at University of Waterloo. It is trying to capture the nature of general human intelligence with algorithms based on how neurons work. Other investors in Google Deep Mind include Peter Thiel and Elon Musk. In exchange for a million pound plus donation to the Computer Sciences Department and Engineering Sciences Department at Oxford, Google Deep Mind gets the benefit of working with some of its leading professors. Engineers at Oxford involved with Google Deep Mind are world leaders in computer vision. The company also purchased Dark Blue Labs, a start up by Oxford computer scientists working on natural language (so that computers and people speaking English or French or Chinese might better understand each other). To the displeasure of the principals at Google Deep Mind, this year Musk warned publicly about the alarming speed with which artificial intelligence is advancing. Musk says he bought into Deep Mind to keep an eye on it.
And what of the Oxford Internet Institute, whose advisory board is chaired by Richard Susskind? It is a graduate school, set up in 2001 to study the social implications of the Internet. No doubt with advice from Susskind Senior, it has broadened its reach to include many things related to the use of Big Data, such as the analysis of issue based conversations on Twitter and "novel crowd sourced intelligence services." There is also interest at OII in teaching computers to understand human languages, statistical modelling and inference, machine learning, etc. It is also one of the founding partners of a brand new entity called the Alan Turing Institute.
The Alan Turing Institute is the British government's attempt to locate itself firmly at Center Ice of Big Data. Other founding partners include four other British universities such as Cambridge (Alan Turing`s alma mater), Edinburgh, Warwick and University College London. According to its website, the Turing Institute "will attract the best data scientists and mathematicians from the UK and across the globe to break new boundaries in how we use big data in a fast moving competitive world." Half of the Institute's start-up funds -- 42 million pounds-- comes from the British government's Engineering and Physical Sciences Research Council. The other half comes from its partners, including the universities, Lloyd's Register, GCHQ, and other businesses. Governments in various countries have put up most of the risk money for the artificial intelligence industry. Now, to speed things up, they are embroiling formerly independent scientists and their employers-- the universities-- in the generation of smart ideas that will generate market applications. Why? Governments argue that this will lead to more and better jobs in the future, which in this case-- the advance of intelligent machine research --gives irony a bad name.
They mainly do it to keep ahead of enemies and frenemies.
GCHQ is the UK's equivalent of the NSA, a leading member of the Five Eyes Consortium which has been shown by documents purloined by Edward Snowden to have secretly gained access to personal, private and foreign government communications world wide. Turing basically invented computing when he worked for GCHQ (then called the Government Code and Cypher School) at Bletchley Park to break the NAZI enigma codes. In those days, the School was under the direction of MI-6. (Please see SMARTS chapter 8 for this story).
The Alan Turing Institute is physically located at the British Library, where Karl Marx was once a Reader dreaming of a post capitalist economy of plenty. There is nothing post-capitalist about the Alan Turing Institute. The new director is Professor Andrew Blake, formerly Microsoft Distinguished Scientist and Laboratory Director of Microsoft Research UK. Last month, when the UK's Minister of Science opened the Institute, he announced that Intel has also agreed to partner.
Next week the Alan Turing Institute will hold a workshop on creating an ethical landscape for the study of Big Data.
Good luck on that.