Tag Archives: AI

AI

The historian Niall Ferguson has made some important points recently on the topic of Artificial Intelligence (AI). Ferguson also published a piece online in Bloomberg. His point was built around the idea that AI should be considered an alien intelligence not to automatically be trusted.

[This isn’t Ferguson, this is my own view.] It is important to remember that the AI phenomenon we’re presently experiencing is that it’s the beginning of an economic bubble. It is a commercial product line constructed for the generation of wealth. It is meant to be a stand-in or augmentation for human beings. This contrivance has the potential for exceeding human abilities in a great many applications in both peace and war. AI will have lightning quick access to the world of information on and off the internet. It will produce both intended and unintended consequences, both of which may be considerable.

AI can be taught to be politically agnostic on one extreme, a political demagogue on the other, or something in between. It can be instructed to commit deception and espionage or angelic kindness. It can be instructed to apply Asimov’s 3 Laws of Robotics or not. Everything depends on the intentions of those in control.

Look around the internet world. It is plagued with troublemakers who are able to escape detection. There is tremendous good available from the internet, obviously, but there are also individuals, criminal enterprises and malevolent governments that are constantly rattling the doorknobs of legitimate websites looking for a way in. Social media has opened up a broad avenue for the efficient delivery of propaganda and swindling to the millions of unwary folk.

AI has a great many applications where it will be of tremendous help to people and organizations. One of the effects of technology advancement is that the time needed for completion of a given task is reduced. Another effect seen since the invention of the wheel is the reduction of labor. Eliminating people from a project reduces costs and increases the concentration of wealth to a few. Understand that cost reductions in business aren’t frequently passed along to consumers. The benefits of cost reductions belong to the stakeholders and will only be directly passed along if absolutely necessary. Cost reductions often manifest as stable pricing over time or a nulling out of inflation, which is a benefit to consumers.

AI will be an honest and moralistic force only if it is instructed to be. Look at the excessively clever application of psychology and show business in advertising. This is a result of fierce competition in the marketplace. Do we really believe that AI will be any different? Given the current state of world affairs, AI will be used as leverage for the transfer of wealth and power from the many to the few. This is the behavior of technological advance.

The promoters of AI will sing its high praises and accuse the doubters of being Luddites. They’ll remind us of the buggy whip and the steam locomotive. In some sense they’ll be right. But our recurring blind spot is with unbridled development. Capitalism as is currently practiced isn’t equipped with much in the way of forward looking moral or existential wariness. It is concerned with the efficient use of capital and distribution of goods. Thoughtful reflection about the future isn’t part of the equation. Greed and desire are the engines of capitalism.

AI will amplify both the best and the worst in us. We must be prepared for the worst because it will come.

Dehumanization of the American Experiment

A few years ago I found myself wandering through the Denver Museum of Nature and Science where I happened upon a robotics exhibition. In terms of the museum arts and sciences it was well conceived and executed, complete with a topical gift shop in the exit. All of the displays were accessible to the public in terms of language or hands-on widgetry. At each hands-on exhibit there stood a determined 5 to 8 year old yanking the controls around in a frantic effort to steer the robotic device away from the wall of the test area while onlookers yawned, waiting their turn. A visitor might have concluded that the purpose of the robot was to become stuck against an obstacle- a task it performed well.

These kinds of future technology exhibits are always popular at the museum. The lead-up to the exhibit is given all of the ballyhoo that the museum could afford. The theme of the exhibit is supercharged with the promise of a brighter tomorrow through the use of snazzy technology. If automobiles can be tied in, so much the better.  It is a celebration of the triumph of technology for the everyman. The subtext was that only by the clever application of technology will we continue to improve our lives. These wonderful robots with their mechanical limbs and primate form would free humans from the dangers and tedium of the work-a-day world.

As I threaded my way through the exhibit I was struck by a sad realization. We’re celebrating the replacement of people with automation. The exhibit was a valentine to all of the entrepreneurs, engineers, investors and vendors who are trying their best to render obsolete much of the remaining workforce. This planned obsolescence has been going for many, many years.

Despite being against our own best interest, we patrons excitedly embrace these “futurama” style exhibitions, perhaps because secretly all of us believe that we will evade the job title of “obsolete”. Absent in the exhibit was a display on what the redundant workers would be doing with their involuntary free time. Fishing or golfing no doubt.

The top-level beneficiaries of robotics are the owners of the factories that make and use them. The driver is that robotics properly done may extend margin growth into the future. A way to overcome foreign competition is by reducing overhead, especially labor costs. Robotics and AI are economic bubbles in the same manner that computers and smart phones have been. The early adopters could enjoy a competitive advantage by the way they use their resources. Profits are unlikely to be channeled into hiring because, well, they’re profiting from the use of robotics. Once automation becomes normalized, there is no going back.

Insider business tip: Healthy companies match labor to the demand for product. More demand, more labor. Increased profits may go towards growth and acquisition, or it may go to the stockholders or to bonuses for management. But rarely if ever a price reduction to the public. If you are making a dandy profit and sales are strong, why hire or reduce prices?

The secondary level beneficiaries will be the consumer who will likely be oblivious to the fact that widget prices have not risen lately. Lower overhead does not automatically result in price savings for the end user. Extra margins will be absorbed by the manufacturer or seller. Just as likely, extra margins may be consumed by the manufacturer in wholesale price negotiations with retailers in the eternal battle for retail shelf space.

Many will offer that the history of man’s use of tools from the stone axe and wheel to AI driven automation is/was inevitable. The ascent of mankind is driven in part by our ability to use tools and develop a command of energy. It is difficult to think of a progressive industrial technology that did not result in the reduction of labor contribution to the overall cost of production. Nobody mourns the loss of the mule team and wagon, steam locomotives, or whale oil. We celebrate obsolescence and we take rapid progress for granted. Technological triumphalism is what we all celebrate.

But we should remind ourselves that there exists a substantial negative aspect of the story of technological progress. It is the very thing it enables: the reduction of labor hours per unit of production. The drive to raise profit margins is relentless, partly because the cost of doing business rises always rises and eats into margins.  Labor costs in particular are always front and center in the mind of business owners.

The situation today is different than when Henry Ford developed his form of mass production. Then there was a smaller population with a significantly larger fraction of people living on farms capable of growing their own food. Many common goods and services were in the hands of local business operators who produced locally and distributed locally. Restrictions on manufacturing and business operations were less onerous than today allowing for greater flexibility in methodology. It may be fair to say that mass production is now widespread and optimized to some degree as a whole. Early automation with just limit switches and relays has given way to microprocessor-controlled process machinery. What is happening presently is the introduction of artificial intelligence (AI). This is the natural progression of technology.

However, we can look a step or two ahead further and ask the question, when will an AI system take over the total management of a factory? When will an AI system have human subordinates? How tight of a leash would we allow an AI system to have on the management of people? The presence of slack in the organization no doubt makes many job descriptions tolerable. What if AI tightened all of the slack in business operations where every half second is accounted for? Would people consent to working for an AI? Companies like Amazon are getting close to this, but there is still human oversight. Extrapolating, it is easy to predict that one day, very quietly, human management will disappear at some level and in its place will be an AI system.

AI has to be taught. Will there be standards of behavior built-in governing how AI interacts with its human subordinates? Will everyone want their companies managed by an AI programmed to have a Jack Welch profile? My god, I hope not.

Another awful thought is the possibility of government and the military run by AI. Let that roll around in your mind for a bit.

There is a need to get back to basic principles here. What is our purpose in life? For most I think it is to love and be loved as well as to participate in some kind of rewarding activity. We all want to be useful and to leave behind some kind of legacy. There is no doubt that the replacement of human labor by AI-driven systems will continue to move forward, encroaching on all of our lives. Ultimately this is driven by a few people at the top who will reap the rewards to the greater concentration of wealth by a few trillionaires. Is concentrated control of limited resources a good thing? Is there any choice?

There is also a large fraction of the population that is not very progressive or forward looking at all. While they enjoy the devices and comforts of advanced technology, they neither understand or care about what is needed to develop a drug or design a new semiconductor chip. Behind our modern civilization is an educated and skilled workforce. However, the US is comprised of many people who are anti-intellectual by nature. This trait has been there all along and will into the future.

In some ways these people are disruptive to the progress and stability of the American experiment and, as of this writing, it isn’t at all clear how this will play out. The USA may well not be a stable enough environment in the future to sustain the continued, very expensive growth of technology. Technological advance requires highly educated workforce who can afford the training to get there. Just to stay even with what we already have, the pipeline of educated people needs to be full.

Forward looking people, the ones who want to sustain our advanced civilization, must step up and be counted or the thing will expire. For all of its problems, the US has nonetheless been a productive incubator of innovation and a great many positive aspects of advanced civilization in the form of a noisy, somewhat chaotic liberal democracy. The goose that laid the golden egg is still alive. Shouldn’t we keep it going?

Meta Making Progress Towards a Science Fiction-Like Dystopia

A piece in the Washington Post by Prashnu Verma appeared reporting progress with Meta’s Cicero artificial intelligence (AI) system. The thrust of the report is that Cicero can play a game called Diplomacy better than humans. The article is worth reading- I know nothing about AI so all I can do is link readers to the article.

Quoting from the Post article-

“Researchers at Meta, Facebook’s parent company, have unveiled an artificial intelligence model, named Cicero after the Roman statesman, that demonstrates skills of negotiation, trickery and forethought. More frequently than not, it wins at Diplomacy, a complex, ruthless strategy game where players forge alliances, craft battle plans and negotiate to conquer a stylized version of Europe.”

Further down …

“It’s a great example of just how much we can fool other human beings,” said Kentaro Toyama, a professor and artificial intelligence expert at the University of Michigan, who read Meta’s paper. “These things are super scary … [and] could be used for evil.”

The nations of the world have civil and criminal laws to discourage and punish people who use their natural intelligence to commit crimes and misdeeds. What about those who use- or unleash- AI to achieve ends that would otherwise be ruled as unethical or even illegal? Pet owners can be held liable for the damage their pets do. Why shouldn’t AI owners have at least the same liability? Could a court order the alteration of an AI’s algorithms in a way that would shut down objectionable or unlawful “behavior”.

If the work product in the application of any intelligence includes action, then where does that leave an AI that can make decisions independently? When could we let it loose to do things that may affect people in novel circumstances? And what kind of ethical responsibility do programmers have in anticipating negative outcomes and acting to arrest them? Lots of questions.

One of the consequences of technological advance has always been the elimination of jobs. That is, getting the same or better results with a lower headcount. It represents cost savings and added margins for an organization. AI will be a valuable tool in the eternal drive for faster-better-cheaper.

AI will almost certainly change many experiences in life. AI systems will manage and replace people in the workplace. It is likely to improve multitasking in many job descriptions, boosting productivity over human counterparts. AI will produce a more effective sales force because the art of persuasion will become much more highly refined. Just what we need- craftier salespersons humping our legs for a sale.

On the positive side, AI has the potential for executing better judgement in many situations. For example, law enforcement could be polished to a point where many errors in judgement can be avoided. This applies across the board in all activities.

AI will also enable criminal intent. The ability to execute crimes will be improved with better judgement, knowledge and fewer mistakes.

Soon, if not already, wars will be guided and fought between AI systems. Cold war type activity could be refined to produce better intelligence and undercover schemes to outwit the other side. Leaders could put AI to use in the darker side of governance. It could be used to keep better track of individuals and information related to them. It could also be used to apply punishment to people without the messy issue of personal morals.

Any dark human activity you can imagine can be made more effective with the application of AI. If it can be tried, it will be tried.