Precarity and Artificial Intelligence: What "HAL" Might Do To "Dave's" Future

Note: the title of this post is a nod to an old, rather slow sci-fi movie with a mind-blowing ending that avoided cheesiness while actually being ahead of its time in many ways... 

This post is a continuation of my series of posts on economic precarity.  As I mentioned in my most recent post in this series, we have been exploring the impact of machine artificial intelligence (AI) on the future of  work, whether that work requires advanced education or not.  But perhaps it might be good to start with a more basic preliminary question: what might be the impact of AI on human life in general?  Of course, the answer to that question is dependent on two factors, namely, the kinds of predictions that are being made concerning the development of AI, and the likelihood of those predictions coming true.  Over the last several years, prognosticators have predicted massive disruptions to human life resulting from the massive and rapid development of AI capabilities.  The tone of these predictions has varied between the optimistic and the dystopian.  Let's limit ourselves to the optimistic for now and ask whether we would want to live in a world in which the most optimistic predictions came true.

One of the more optimistic points of view can be found in a book published in 2021 titled, AI 2041:

Ten Visions for Our Future, by Kai-Fu Lee and Chen Qiufan.  Dr. Kai-Fu Lee holds a PhD from Carnegie Mellon University and has founded or led a number of tech companies as well as doing extensive research and writing in the field of artificial intelligence.  Chen Qiufan is Chinese science fiction writer who formerly worked for tech companies Google and Baidu before launching into a full-time creative career.  AI 2041 is a multifaceted picture of Kai-Fu Lee's predictions of the evolution of AI capabilities from now to the year 2041, combined with Chen Qiufan's short stories portraying fictional settings in which each of these predictions comes true.  Among the things which Dr. Lee believes we are most likely to encounter are the following:

• The use of deep learning and big data paired with social media to guide customers of financial products into decisions and lifestyles which have the least risk of adverse outcomes and the greatest chance of net benefit as calculated by an AI objective function.  (See the story "The Golden Elephant.")
• The use of natural language processing and GPT as tools for creating customized virtual "teachers" for children.  (See the story "Twin Sparrows.")
• The use of AI tools for the rapid analysis of pathogens and the rapid development of drugs for emerging new diseases, as well as the use of automation in management of epidemics and pandemics.  (See the story "Contactless Love.")
• The displacement of skilled manual laborers by AI, and the use of AI to create virtual solutions for this displacement which return some sense of purpose to workers who have lost their jobs.  (See the story "The Job Savior.")
• The ways people cope with the likely displacements and disruptions which will be experienced by societies in which having one's basic needs met becomes decoupled from having to work to earn a living. (See the story "Dreaming of Plenitude.") 

Note that I have listed only five of the ten possible scenarios sketched by Dr. Lee.  However, these five are most relevant to the topic of today's post.  It is already becoming possible to design AI-powered virtual "life coaches" to guide people in their life decisions.  (In fact, if you really want to let your bloody smartphone tell you how to run your life, you can find apps here, here, and here for starters.)  However, when using these apps, one must remember that at their heart they are simply machines for optimizing objective functions which have been designed by humans and which have been tuned by massive amounts of human-supplied training data.  Thus these "coaches" will be only as smart (or as stupid) as the mass of humanity.  And they can be made to encode and enshrine human prejudices, an outcome which is especially likely whenever decisions involving money or social power are involved.  This is illustrated in the story "The Golden Elephant."  (For a harder-edged, more pessimistic view of this sort of AI application, please check out the short story "The Perfect Match" by Ken Liu.)

The use of AI tools in medicine for discovery of pathogen structure and rapid drug development is a fine example of the emerging use of machine implementation of multi-objective function optimization.  I truly have nothing but praise for this sort of application, as it has saved countless lives in the last half decade.  For instance, this sort of technology was instrumental in the rapid development of safe and effective COVID vaccines.  However, when we get to the use of AI to replace the kind of skilled labor that has historically depended on the development of human cognitive capabilities, I think we're headed for trouble.  Consider the case of teaching children, for instance, as exemplified by Chen Qiufan's short story "Twin Sparrows."  Teaching in modern First World societies has evolved into the delivery of a standardized curriculum by means of standardized methods to children, and the evaluation of the learning of these children by means of standardized tests.  

Now I know a little about teaching children, as I volunteered for a few years to be an after-school math coach.  And I can tell you that teaching arithmetic to one or a few children requires more than just knowing arithmetic.  It also involves emotional intelligence and the skill of careful observation as well as a certain amount of case-by-case creativity.  We must ask whether these things can be captured by an AI application that has been "optimized" to maximize learning.  How does one measure things like student engagement?  For instance, do we write some polynomial regression function in which one of the terms stands for whether the kid's pupils are dilated, another term stands for whether a kid's eyes are open and looking at the teacher or whether they're closed, another term captures whether a kid is sitting quietly or throwing a fit, etc.?  And what happens when we move beyond a standard curriculum?  How, for instance, do you make an AI "virtual" art teacher?

I won't attempt to answer these questions here, although I will mention that China has already begun to deploy AI in primary school education, as noted in the 2020 Nesta article titled, "The Future of the Classroom? China’s experience of AI in education" and the 2019 article "Artificial intelligence and education in China," which is unfortunately behind a Taylor and Francis paywall.  It will be interesting to see comprehensive, multi-year studies which document whether the use of AI in education is actually living up to its promise.  

But let's say that the deployment of AI in education really does turn out to be effective.  What happens to the human teachers in such a case?  Kai-Fu Lee says in AI 2041 that teachers will still be needed to be confronters, coaches, and comforters.  In fact, this seems to be a rather stock answer given whenever the potential massive occupational disruptions promised by the widespread deployment of AI are mentioned.  We are told that when jobs that formerly required powers of observation, quick assessment, logical reasoning, computational or motor skills, or memorization are taken over by AI, the newly-displaced workers can be retrained as "compassionate caregivers."  But it might be good to confront the fact that the widespread deployment of AI under an optimistic scenario would certainly mean the de-skilling of large numbers of people.  What possibly unforeseen effects would this de-skilling have on the displaced workers even if they were retrained as "compassionate caregivers?"

Consider, for instance, what might happen to London cab drivers if they were replaced by self-driving taxis.  To become a London taxi driver, a person must memorize a huge amount of London metro local geography, then pass a special test administered by the British government.  (From what I hear, you can't cheat on the test by using a GPS!)  All that memorization (especially visual memorization of London streets and intersections) induces strong development of key regions of the brains of aspiring London taxi drivers.  If this challenge is taken away from a London cabbie, he or she will lose that brain development.  Consider also the personnel who comprise flight crews of airliners.  Up to the 1960's, one of the positions on the flight deck of an airliner was the navigator.  But the navigator position was eliminated by autopilots.  So flight crews shrank from four to three people.  But then, further advances in automation eliminated the position of flight engineer.  So now flight crews consist of only two people.  What development was lost in the brains of the navigators when they were replaced by machines?  (What navigational feats are humans capable of when those humans are pushed to their cognitive limits?  Consider for instance how the peoples of Oceania learned to sail between their islands reliably and successfully without needing maps or a compass.)

AI has eliminated not only aircraft navigators and flight engineers, but an increasing number of other degreed professionals including medical radiologists, as well as receptionists, telephone operators, fast-food cooks, waiters, and waitresses.  AI "expert systems" are threatening the jobs of an increasing number of skilled, educated technical professionals, as noted here and here, for instance.  An increasing number of news stories are documenting the ongoing erosion of human labor markets by AI.  It must be asked what will happen to people whose jobs required the development of hard cognitive skills when those skills are replaced by AI.  Preliminary answers to that question are not encouraging.  For instance, the British Journal of Medicine published a 2018 article titled, "Intellectual engagement and cognitive ability in later life (the “use it or lose it” conjecture): longitudinal, prospective study," in which the authors concluded that lifelong intellectual engagement helps to prevent cognitive decline later in life.  There is also a 2017 article published in the Swiss Medical Weekly whose authors concluded that "low education and cognitive inactivity constitute major risk factors for dementia."  In other words, by ceding to AI the hard cognitive challenges which have traditionally been the hallmark of many kinds of paying work, we may well be at risk of turning ourselves into a society of de-skilled idiots.

Ahh, but there's more.  Let's consider the obvious fact that when AI takes over a job, one or more humans is thrown out of work.  Let's consider the response of various politicians to this fact.  For instance, let's consider the rhetoric spouted by crooks like Donald Trump and other Republican Party politicians (as well as their millions of adoring fans) in the run-up to the 2016 election.  Let's also consider the "scholarly" articles, ethnographic studies and books such as Hillbilly Elegy which sought to "explain" the Trump phenomenon.  One of the key assertions of the Trump crowd in 2016 was that the reason why the white American working class was becoming increasingly poor was the threat posed by immigrants (especially dark-skinned immigrants) taking jobs away from "real" Americans.  Thus America needed to build walls - made both of barbed wire and cement, and of policies and legislation -  in order to keep the great unwashed from stealing what "rightfully" belongs to America.  In other words, one of the biggest drivers of the growth of Trumpism was the loss of jobs and income among the white American working class.  But if concern about job losses was really so bloody important to the architects of Trumpism, why is it that they did not utter a single word in protest against the threat to jobs posed by the deployment of AI?  Why is it that NO ONE in the Rethuglican Party nowadays has anything bad (or even cautionary) to say about the use of AI by American businesses?  The silence of the Rethuglicans regarding the disruptions of AI can be explained quite simply.  AI helps business owners increase profits while reducing labor costs.  Thus AI helps the rich get richer.  Also, Trumpism is not and never was about bringing jobs back to the "working class".  It was rather always an expression of collective narcissism.  Thus all the talk about jobs, like all the rest of the rhetoric of the American Right, was and is utter crap.

To be sure, we do need to start having urgent conversations, both locally and on a wider scale, regarding the deployment of machine artificial intelligence in society.  Such conversations need to ask what AI can reasonably be expected to be able to do, as well as asking whether we really need machines to do what AI is promised to do.  If we decide that it is actually in our best interest to continue the massive development and deployment of AI, we need to figure out how to do this in such a way that we maximize the benefits of AI while minimizing our exposure to the potential downsides and negative externalities of AI.  Lastly, we need to start asking whether it might make sense to establish a basic universal income and other social structures which allow the people in our societies to develop their full human potential even in an era of the expanding use of AI.

Precarity and Artificial Intelligence: A Four-Wheeled Reason to be Skeptical about AI Optimism

The most recent post in my series on economic precarity hinted that the wildly optimistic claims of what artificial intelligence can do or is about to be able to do may be a bit overblown.  A case in point just surfaced this week: the Tesla Corporation (and its CEO Elon Musk in particular) are now being investigated by Federal prosecutors about claims made by Musk that Tesla's "self-driving car" AI technology has actually produced cars that drive themselves without any human input.  It seems this claim is not quite true, as "hundreds of crashes and dozens of fatalities" have proven over the last few years.  Musk may soon find himself the target of State-sponsored vengeance - a vengeance carried out by human prosecutors, plaintiffs, judges, and juries instead of robots.  They may optimize their "objective function" to return a guilty verdict.  Could this be the start of a rocky road for Musk ... ?

Precarity and Artificial Intelligence: Review of Objective Functions, and A Contrarian Perspective

This post is a continuation of my series of posts on economic precarity.  As I mentioned in recent posts in this series, we have been exploring the subject of the educated precariat - that is, those people in the early 21st century who have obtained either bachelors or more advanced graduate degrees from a college or university, yet who cannot find stable work in their chosen profession.  However, the most recent previous post in this series began to explore the impact of machine artificial intelligence (AI) on the future of all work, whether that work requires advanced education or not.  

In my my most recent post in this series, I wrote that the greatest potential for the disruption of the future of work through machine AI lies in the development of machine AI tools that can tackle the increasingly complex  tasks that are normally associated with human (or at least highly developed animal) intelligence.  Such tasks include machine vision (including recognizing a human face or an animal), natural language processing, construction of buildings, navigating physically complex unstructured and random environments (such as forests), and optimization of problems with multiple objectives requiring multiple objective functions to model.  Today I'd like to amend that statement by saying that there is another potentially massive disruptive impact of machine AI on the future of work, namely, the ways in which the wide deployment of machine AI in a society might condition and change the humans in that society.  I'll have more to say on that subject in another post.  What I'd like to do right now is to take another look at mathematical objective functions and their place in the implementation of machine AI.  But even before that, let's review the two main types of applications we are talking about when we talk about "machine AI".  Also, let me warn you that today's post will move rather deep into geek territory.  I'll try to have some mercy.

As I mentioned in the most recent post in this series, AI applications can be broken down into two main categories.  The first category consists of the automation of repetitive or mundane tasks or processes in order to ensure that these processes take place at the proper rate and speed and thus yield the appropriate steady-state or final outcome.  This sort of AI has been around for a very long time and first came into existence in entirely mechanical systems such as steam engines with mechanical governors that regulated the speed of the engines.  An example of a mid-20th century electromechanical control system is the autopilot with mechanical gyroscopes and accelerometers and simple electronic computers which was invented to guide airliners and early guided missiles during long-distance flights.  Other early electronic examples include the programmable logic controllers (PLC's) which were developed in the 1960's to regulate assembly line processes in industrial plants.  In his 2022 paper titled "The two kinds of artificial intelligence, or how not to confuse objects and subjects," Cambridge Professor Alan F. Blackwell characterizes these systems as servomechanisms, which Merriam-Webster defines as "an automatic device for controlling large amounts of power by means of very small amounts of power and automatically correcting the performance of a mechanism" and which Blackwell himself defines as "...any kind of device that 'observes' the world in some way, 'acts' on the world, and can 'decide' to 'behave' in different ways as determined by what it observes."  

The construction (and hence performance) of servomechanisms can be increasingly complex as the number and type of processes regulated by the servomechanisms increases, but that does not mean that the servomechanisms possess any real native intelligence.  Consider, for instance, a very simple servomechanism such as the thermostat from a heating system in a house built during the 1950's.  Such a thermostat would regulate the timing and duration of the burning of a fuel in a heating furnace, and would most likely consist of a simple switch with a movable switch contact and a stationary contact with the movable contact attached to a bimetallic strip.  Because the shape of the bimetallic strip is regulated by the temperature of the air, when the air temperature drops, the movable switch contact eventually touches the stationary contact, closing the switch and turning on the flow of fuel to the furnace.  We can say that the thermostat "decides" when the furnace turns on or off, but that's all this thermostat can "decide" to do.  You certainly wouldn't want to rely on the thermostat to help you decide what movie to watch with your spouse on a weekend!  Servomechanisms are what Blackwell calls "objective AI" which "measures the world, acts according to some mathematical principles, and may indeed be very complex, even unpredictable, but there is no point at which it needs to be considered a subjective intelligent agent."  In other words, all a servomechanism can do is to mechanically or electronically regulate a physical process on the basis of process measurements provided to the controller by means of physical sensors that sense a process variable.  It can't think like humans do.

The second type of AI is designed to make value judgments about the world in order to predict how the world (or some small subset of the world) will evolve.  In the most optimistic cases, this AI uses these value judgments to generate the most appropriate response to the world which is supposedly evolving according to prediction.  But is this really a native intelligence created by humans, yet now embodied in a machine and existing independently of humans?  A possible answer to that question can be found in another paper written by Blackwell and published in 2019, titled, "Objective functions: (In)humanity and inequity in artificial intelligence."

The value judgments and predictions made by the second type of AI are made by means of objective functions. These objective functions are mathematical abstractions consisting of functions of several independent variables.  Each of the independent variables represents an independently controllable parameter of the problem.  If the purpose of the objective function is to predict the numerical value of an outcome based on historical values of independent input variables, then optimizing the function means making sure that for a given set of historical inputs, the objective function yields an output value that is as close as possible to the historical outcome associated with the particular historical inputs. This ensures that for any set of future possible inputs, the objective function will accurately predict the value of the output.  Two levels of objective functions are needed: the first level, which makes a guess of the value of an output based on certain values of inputs, then a second supervisory level which evaluates how close each guess is to a set of historical output values based on corresponding sets of input values.  The output of this second supervisory objective function is used to adjust the weights (in the case of a polynomial function, the coefficients) of the primary objective function in order to produce better guesses of the output value. 

Objective functions are mathematical expressions; hence, the second type of AI is a primarily mathematical problem which just happens to be solved by means of digital computers.  This also includes the implementation of multi-objective optimization, which is really just another mathematical problem even though it is implemented by machines.  Thus, the second type of AI is really just another expression of human intelligence.  This is seen not only in the development of the objective functions themselves, but also in the training of the supervisory objective function to recognize how close the output of the primary objective function is to the a value that actually reflects reality.  This training takes place by several means, including supervised learning (in which humans label all the training data), and partially-supervised and unsupervised learning (in which the training data is out there, but instead of it being labeled, a human still has to create the algorithms by which the machine processes the training data).  

An example that illustrates what we have been considering is the development of large language models (LLM's) such as ChatGPT which predict text strings based on inputs by a human being.  A very, very, very simple model for these AI implementations is that they consist of objective functions that guess the probability of the next word, phrase, sentence, or paragraph in a string on the basis of what a human has typed into an interface.  These AI implementations must be trained using data input by human beings so that they can calibrate their objective functions to reduce the likelihood of wrong guesses.  Cases like these lead scientists such as Alan Blackwell to conclude that the second type of AI is not really a separate "intelligence" per se, but rather the embodiment and disguising of what is actually human intelligence, reflected back to humans through the intermediary of machines.  The calibration of the objective functions of these AI deployments (or, if you will, the training of this AI) is performed by you every time you type a text message on your smartphone.  For instance, you start by typing "Hello Jo [the phone suggests "Jo", "John", "Joe", but the person you're texting is actually named "Joshiro", so as you type, your phone keeps making wrong guesses like "Josh", "Joshua", and "Josh's" but you keep typing until you've finished "Joshiro"]. You continue with "I'm at the gym right now, but I forgot my judo white belt [the phone guesses almost everything even though you misspelled "at" as "st" and the phone auto-corrected.  However, the phone chokes when you start typing "judo" so you have to manually type that yourself].  You finish with "Can you grab it out of my closet?"  The next time you text anyone whose first name starts with the letters "Jo", your phone will be "trained" to think you are texting Joshiro about something related to judo - or more accurately, the generative LLM in your phone will have determined that there is a statistically higher likelihood that your message will contain the words "Joshiro" and "judo".  Your phone's LLM is thus "trained" every time you correct one of its wrong guesses when you type a text.

Of course, the longer the predicted phrase or sentence or paragraph the AI is supposed to return, the more training data is required.  The boast of the developers of large language models and other similar AI implementations is that given enough training data and a sufficiently complex statistical objective function, they can develop AI that can accurately return the correct response to any human input.  This unfortunately leads to an unavoidable conclusion of the second type of AI: the assumption that the universe, reality, and life itself are all deterministic (thus there is no free will in the universe).  Why? Because the kind of intelligence that can accurately predict how the universe and everything in it will evolve and thus generate the most appropriate local response to the moment-by-moment evolution of your particular corner of the universe can always be modeled by an appropriately elaborate statistical objective function trained on an appropriately huge set of training data.  In other words, given enough data, a statistical objective function can be derived which accurately predicts that your spouse's sneeze at the dinner table tonight will provoke an argument which ends with you sleeping on the couch tomorrow night, and that this will lead to the invention of a new technology later in the week that causes the stock market of a certain country to crash, with the result that a baby will cry on a dark and stormy night five days from now, and that this cry will enable a computer to predict all the words that will be in the novel you sit down to write a week from today...  This is my rather facetious illustration of the "generative AI" of chatGPT and similar inventions.  The fallacy of determinism is that life, the Universe, and reality itself are full of phenomena and problems that can't easily be modeled by mathematics.  Scientists call them "wicked problems."  Thus the claims made about the second kind of AI may be overblown - especially as long as the implementation of this second kind of AI remains primarily dependent on the construction and optimization of appropriately complex statistical objective functions.

Yet it can't be denied that the second type of AI is causing some profound changes to the world in which we live, and even the first type of AI - the implementation of servomechanisms, or the science of cybernetics - has had a profound effect.  The effects of both types of AI to date - especially on the world of work - will be the subject of the next post in this series.

P.S. Although I am a technical professional with a baccalaureate and a master's degree in a STEM discipline, I am most definitely NOT an AI expert.  Feel free to take what I have written with a grain of salt - YMMV. 

P.P.S. For more commentary on ChatGPT and other LLM's, feel free to check out a 2021 paper titled, "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" by Emily Bender and others.  The authors of this paper were Google engineers who were fired by Google for saying things that their bosses didn't want to hear regarding LLM's.   Oh, the potential dangers of writing things that give people in power a case of heartburn...

 

Precarity and Artificial Intelligence: The Foundations of Modern AI

This post is a continuation of my series of posts on economic precarity.  As I mentioned in recent posts in this series, we have been exploring the subject of the educated precariat - that is, those people in the early 21st century who have obtained either bachelors or more advanced graduate degrees from a college or university, yet who cannot find stable work in their chosen profession.  Today's post, however, will begin to explore a particular emerging impact to employment for everyone, whether formally educated or not.  That impact is the impact of machine artificial intelligence on the future of work.

As I mentioned in a previous post, 

Labor casualization has been part of a larger tactical aim to reduce labor costs by reducing the number of laborers...This reduction of the total number of laborers can be achieved by replacing employees with machines.  That replacement has been occurring from the beginning of the Industrial Revolution onward, but in the last two or three decades it has accelerated greatly due to advances in artificial intelligence (AI).  A long-standing motive behind the recent massive investments in research in artificial intelligence is the desire by many of the world's richest people to eliminate the costs of relying on humans by replacing human laborers with automation.

So it is natural to ask what sort of world is emerging as the result of the use of increasingly sophisticated AI in our present economy.  Here we need to be careful, due to the number of shrill voices shouting either wildly positive or frighteningly negative predictions about the likely impacts of AI.  I think we need to ask the following questions:

• First, what exactly is artificial machine intelligence?  What is the theoretical basis of AI?  How does it work? ...

Today we'll start trying to answer the questions stated above.   And at the outset, I must state clearly that I am not an AI expert, although my technical education has exposed me in a rudimentary way to many of the concepts that will be mentioned in our discussion of AI.

First, let's paint a picture.  One of the original motives for trying to invent intelligent machines was the desire for machines that would reliably do the kind of mundane tasks that humans find to be distasteful or unpleasantly difficult.  This desire actually has a very long history, but was popularized in the science fiction of the mid-20th century.  Think of a kid from the 1950's or 1960's who wished he had a robot that could vacuum the living room carpet or take out the garbage or do homework or shovel snow out of the driveway so that the kid could play without being bothered by parents demanding that the kid himself do the tasks listed above.  What kind of "brain" would the robot need in order to know what tasks needed to be performed and when would that brain know the tasks had been performed to acceptable standards?

The question of the kind of "brain" required was solved by the invention of the first programmable all-electronic digital computer during World War Two.  This computer was itself an evolution of principles implemented in previous mechanical and electromechanical computers.  Once engineers developed digital computers with onboard memory storage, these computers became capable of automation of tasks that had been formerly automated by more crude mechanical and electromechanical means such as relays.  Thus the 1950's saw the emergence of digital control systems for automation of chemical processes at refineries; the 1950's and 1960's saw the emergence of computer-assisted or computer-based navigation for ships, aircraft, missiles, and spacecraft; and the late 1960's saw the emergence of programmable logic controllers (PLC's) for automation of factory processes at industrial assembly plants.

The digital electronic automation systems that have been developed from the 1950's onward have thus formed a key component of the development of modern machine AI.  But another key component consists of the principles by which these systems achieve their particular objectives.  These principles are the principles of mathematical optimization, and they also have a rather long history.

Mathematical optimization is the collection of techniques and methods for finding the maximum or minimum value of a mathematical function of one or more independent variables.  Some of the earliest methods of mathematical optimization were based on calculus.  More complex methods include such things as numerical methods for solving nonlinear differential equations.  These methods were only possible to implement easily once electronic digital computers became available.

The first step in optimizing real-world problems consists of turning a real-world problem into a mathematical abstraction consisting of a function of several independent variables.  Each of the independent variables represents an independently controllable parameter of the problem.  Then optimization techniques are used to find the desired maximum or minimum value of the function.  To put it another way,

"Optimization is the act of obtaining the best result under given circumstances.  In design, construction, and maintenance of any engineering system, engineers have to take (sic) many...decisions.  The ultimate goal of all such decisions is either to minimize the effort required or to maximize the desired benefit.  Since the effort required or the benefit desired in any practical situation can be expressed as a function of certain decision variables, optimization can be defined as the process of finding the conditions that give the maximum or minimum value of a function."  - Engineering Optimization: Theory and Practice, Rao, John Wiley and Sons, Inc., 2009

The function to be optimized is called the objective function.  When we optimize the objective function, we are also interested in finding those values of the independent variables which produce the desired function maximum or minimum value.  These values represent the amount of various inputs required to get the desired optimum output from a situation represented by the objective function.  

A simple case of optimization would be figuring out how to catch up with a separate moving object in the shortest amount of time if you started from an arbitrary starting position.  To use optimization techniques, you'd turn this problem into an objective function and then use calculus to find the minimum value of the objective function.  Since the velocity and acceleration are the independent variables of interest, you'd want to know the precise values of these (in both magnitude and direction) which would minimize the value of the objective function.  Note that for simple trajectories or paths of only two dimensions, adult humans tend to be able to do this automatically and intuitively - but young kids, not so much.  Try playing tag with a five or six-year-old kid, and you will see what I mean.  The kid won't be able to grasp your acceleration from observing you, so he will run to where he sees you are at the moment he sees you instead of anticipating where you'll end up.  Of course, once kids get to the age of ten or so, they're more than likely to catch you in any game of tag if you yourself are very much older than 30!

The easiest AI problems are those that can most easily be turned into mathematically precise objective functions with only one output variable.  Examples of such problems include the following: reliably hitting a target with a missile, winning a board game in the smallest number of moves, traveling reliably between planets, simple linear regression, regulating the speed or rate of industrial or chemical processes, and control of HVAC and power systems in buildings in order to optimize interior climate, lighting, and comfort.

Harder AI problems include machine vision (including recognizing a human face or an animal), natural language processing, construction of buildings, navigating physically complex unstructured and random environments (such as forests), and optimization of problems with multiple objectives requiring multiple objective functions to model.  The machine vision and natural language processing problems are harder because they require the use of logistical regression functions as objective functions, and in order to accurately assign the appropriate "weights" to each of the variables of these objective functions, the AI which implements these functions needs massive amounts of training data.  However, these and other harder problems are now being solved with increasing effectiveness through technologies such as deep learning and other advanced techniques of machine learning.  It is in the tackling of these harder problems that AI has the greatest potential to disrupt the future of work, especially of cognitively demanding work that formerly only humans could do.  In order to assess the potential magnitude and likelihood of this disruption, we will need to examine the following factors:

• The current state of the art of machine learning
• The current state of the art of designing objective functions
• And the current state of the art of multi-objective mathematical optimization.

I'll try tackling these questions in the next post in this series.

Precarity, Late Capitalism, And Artificial Intelligence: Pinocchio's Mischief

This post is a continuation of my series of posts on economic precarity.  As I mentioned in recent posts in this series, we have been exploring the subject of the educated precariat - that is, those people in the early 21st century who have obtained either bachelors or more advanced graduate degrees from a college or university, yet who cannot find stable work in their chosen profession.  The two most recent previous posts in this series discussed the fact that there are now more college graduates being produced in our society than there are jobs into which to plug those graduates.  The most recent post discussed why this is the case.  As I wrote last week, 

"...the decline in opportunities for college graduates (along with everyone else) is correlated with the rise in the concentration of economic power in the hands of an ever-shrinking elite.  In fact, I will go even farther and assert that the decline in stable employment for college graduates (even those with technical professional degrees) is a direct outcome of the concentration of economic power at the top of society.

Consider the fact that as of 2015, "America's 20 wealthiest people - a group that could fit comfortably in one single Gulfstream G650 luxury jet - now own more wealth than the bottom half of the American population combined..."  These people therefore have an enormous amount of economic and political clout.  And they have used (and continue to use) that clout in order to turn the American economy into a machine whose sole function is to make them as rich as possible.  The increase in precarity, the casualization of increasing types of employment, and the increasing use of task automation and artificial intelligence are typical of the strategies which these wealthy and powerful people have deployed in order to maximize the wealth they can extract from the American economy while minimizing the amount of wealth they give to the rest of us.  The aggressive expansion of the "gig" economy is another such strategy..."

A basic strategic aim in capitalism is that business owners should maximize profit.  A basic tactic for the achievement of this aim is to maximize profit per unit of goods sold by lowering the cost of production for each unit of goods sold.  Lowering costs can be achieved by attacking the cost of materials, capital machinery, energy, and labor.  In the limit, at the extreme of optimization, this leads to extremely flimsy goods sold for extremely high prices, goods that are produced by extremely poor laborers.

The labor part of this tactic is what we have been discussing in our consideration of precarity.  By making employment casual and temporary, with no fixed covenant between businesses and laborers and no benefits (other than a wage) granted to laborers, businesses have succeeded in driving down the cost of labor.  As mentioned in last week's post, that pressure on labor costs has reached even technical professions requiring a baccalaureate degree or above.  This is leading to an increasingly unsustainable situation in which, for instance, you might spend more than $40,000 to earn a four-year engineering degree - only to find yourself working for an engineering temp agency after graduation!

Labor casualization has been part of a larger tactical aim to reduce labor costs by reducing the number of laborers.  If you're the CEO of a large company, the progression of this tactic can be sketched as follows: First, destroy any expectation of stable employment or decent wages among your labor pool.  Then, reduce the actual number of laborers you use.  This reduction of the total number of laborers can occur by a number of means (including working employees to death by giving each employee the amount of work that should normally be handled by two or three such employees).  It can of course also be achieved by replacing employees with machines.  That replacement has been occurring from the beginning of the Industrial Revolution onward, but in the last two or three decades it has accelerated greatly due to advances in artificial intelligence (AI).  A long-standing motive behind the recent massive investments in research in artificial intelligence is the desire by many of the world's richest people to eliminate the costs of relying on humans by replacing human laborers with automation.

So it is natural to ask what sort of world is emerging as the result of the use of increasingly sophisticated AI in our present economy.  Here we need to be careful, due to the number of shrill voices shouting either wildly positive or frighteningly negative predictions about the likely impacts of AI.  I think we need to ask the following questions:

• First, what exactly is artificial machine intelligence?  What is the theoretical basis of AI?  How does it work?
• What can AI do and not do?
• What countries are at the forefront of AI deployment in their societies?
• How will AI capabilities likely evolve over the next few decades?
• What effects might AI have on human life and human societies over the next few decades?
• How will AI affect the world of work over the next few decades?

The next few posts in this series will attempt to tackle these questions.  I must warn you that what you'll get in those posts are merely my guesses at an answer.  However, because I want the guesses to be educated guesses, I'm going to need to do some research.  So these guesses might be slow in coming.