Almost 70 years ago, precisely on December 28, 1948, a learned monk, with serious philosophical interests and well-informed about the most recent developments in the techno-scientific research of his time, Father Dominique Dubarle, sent a letter to the widely circulated Parisian newspaper Le Monde, to comment on Norbert Wiener’s (largely technical) book Cybernetics, which had just been published; and, of course, he had read it. 1 Father Dubarle did not leave his name in history like Wiener did. However, his commentary at the time has acquired a contemporary interest:
…Thus, one of the most fascinating prospects that emerge is the possibility of logically regulating human affairs, and more specifically those concerning human communities that exhibit a certain statistical regularity, such as the human formation of public opinion. Could one, therefore, imagine a machine that would collect any kind of information, such as information regarding production and the market; and then determine what the most likely development of the situation would be, according to the average psychology of human beings, and the quantities that can be measured in each specific case.
Is it possible to conceive of a state machine that would encompass all systems of political decision-making, under a regime of many states distributed across the surface of the earth, or under an obviously much simpler regime of a government for the entire planet? At present, nothing prevents us from imagining such a thing. We can dream of the era when it will be possible for a machine à gouverner (machine for governing) to supplement – for better or for worse – the current obvious inadequacy of the human brain when it deals with the usual mechanism of politics.Undoubtedly, human realities do not admit a clear and certain determination, as is the case with the numerical data of mathematics. The only thing they admit is the determination of their probabilities. A machine, therefore, that deals with these processes, must rather use a probabilistic way of thinking than a deterministic one, as is the case, for example, in modern computing machines.
This fact makes its operation even more complicated, but not impossible. The prediction machine that determines the effectiveness of the anti-aircraft gun is a machine of this type.Theoretically, predicting the most likely time of an event is not possible. Nor is it possible to determine the most favorable decision, at least within certain limits. It can be assumed that the possibility of existence of machines such as the chess machine ensures this fact. And this happens because human processes that constitute the subject of governance can be compared with “Games”, in the sense in which von Neumann studied them mathematically.
Although these “Games” are determined by an incomplete set of rules, there are other “Games” with a very large number of players, where the data are extremely complex. For the machines to govern the state, it will be determined as the best-informed player at each particular level; and the state will be the sole supreme coordinator of all individual decisions.The advantages will be enormous. If, however, they are acquired in a scientific manner, they will allow the state to defeat any player, except itself, in a human game, under any conditions, by confronting them with the following dilemma: either suffer immediate destruction, or cooperate according to plan. These will be the consequences of the game itself without the use of external violence. Lovers of an ideal world thus have something real to dream about.
Dubarle’s inspiration for a governing machine could then (after such) be discussed in the salons of idle dreamers; otherwise it would be laughable. Dubarle himself did not exclude its invention; however, he did not see it as “available very quickly”.
Wiener, for his part, commenting on Dubarle’s inspiration, was only concerned that the machine a gouverner might fall into the hands of “bad people”: the 3rd world war (the so-called “cold”) had already begun…
…
The machine to govern of Father Dubarle is not frightening because there is some danger that it will achieve autonomous control of humanity… The real danger it presents is the entirely different danger that such machines, although harmless in themselves, can be used by a human being or by a group of human beings to increase their control over the rest of humanity, or by political leaders in their attempt to control the population, not through these machines themselves but through a political technique so narrow and indifferent to human capabilities that one could say that in reality this policy has been designed by a machine.
The great inability of this machine – the inability that saves us from its dominion – is that it still cannot take into account the vast field of probability that characterizes the human condition. The dominion of the machine presupposes a society in its final stages of increased entropy, where probabilities are negligible and there are no statistical differences between individuals. Fortunately, we have not yet reached this stage.
…
However, seventy or eighty years is not “too late” – if human History is the measure. In the most recent pre-election period in the US, in 2016, there were not only two candidates, Trump and Clinton. There was (or, more correctly, there was an attempt for there to be) a third one. Watson. Watson is a machine, a creation of IBM’s engineers, and is considered one of the most advanced constructed and operational systems to date that could be called artificial intelligence.
On the website of supporters of Watson’s presidential candidacy it was written:
The Watson Foundation 2016 is an independent organization created to support the candidacy of artificial intelligence known as Watson for the presidency of the United States of America. We believe that Watson’s unique abilities to access information and make informed and transparent decisions make him the ideal candidate for the responsibilities required of a president.
Watson is a system of computational processes used to provide answers to questions asked in natural language, developed by IBM for the Jeopardy! quiz show. Watson combines information from a variety of sources, through multiple terabytes of data that are used as the basis for generating answers. The more information Watson can consume, the more substantiated its ability to answer becomes. It is also capable of receiving information from any source, allowing it the possibility to analyze different perspectives and political agendas for any given topic.
Watson represents a milestone in machine learning, because it was designed to compete against humans, using natural language, processing with accuracy in answering but also with speed. It understands a question, uses the basic information elements of the question to analyze a huge volume of data, and retrieve the answers that are most likely to be correct. This is a task that all politicians do on a daily basis, including the president, and can be made more effective and accurate by artificial intelligence.
Watson is an application of the DeepQA system, a massively parallel probabilistic architecture designed to handle vast amounts of data and generate the best possible answers to any given question. This data can be of any type, from trivia to geopolitical issues, and even healthcare topics. Since 2013, it has been used to assist in decision-making for clinical care of lung cancer patients at Memorial Sloan-Kettering Cancer Center, where 90% of nurses who use the system follow its guidance.
What makes Watson unique is his ability to relate to people. Not only does he answer verbally, but he also has a visual representation that shows his current state each time. Just as people have facial expressions to express their feelings, Watson changes his visual form to express his level of certainty in each specific answer.
Wouldn’t it be better for the country if all politicians were so transparent?Watson can analyze trends in employment, markets, interest rates, education, poverty, crime, and taxation, in order to determine which actions are most suitable to accelerate investment in the country’s future. We believe that investing in youth is the key to America’s success in the coming century, and that significant improvements can be made in many areas through federal budgeting and tax reform. Some of the actions that could make the country competitive on a global scale again include:
– National provision of healthcare services to all
– Free access to higher education
– Combating homelessness
– Legalization and regulation of personal drug use.
…
It is, perhaps, a shame that Watson did not choose on his own to run as a candidate in the last U.S. presidential elections, against both Clinton and Trump: it would have been interesting to learn that, by analyzing huge amounts of data concerning American society (the voters), he concluded in favor of a left-liberal agenda such as the one above! But no, this agenda was a human choice, made by those supporting his candidacy.
Whether due to this political direction, or because the inspiration of Dubarle has not yet matured, Watson would not have fared well as a U.S. presidential candidate in 2016. In fact, his candidacy went largely unnoticed outside certain technophile circles. One of them, Parag Khanna, wrote after the elections, on January 23, 2017, on his blog:
The best candidate for the position of president of the USA ultimately failed to participate in the November 1st elections. It was Watson, IBM’s conscious computer. Watson would have been a president with intelligence and integrity, composure and dedication, working 24 hours a day to find solutions to the most difficult domestic and international problems. In complete contrast to any human leader, Watson is not a political animal or a personality that sells pies, but a political machine built from various materials…. The American political system needs much less of what Hillary Clinton and Donald Trump are, and much more Watson.
Western societies are collapsing into populist movements that promise, basically, nostalgia. It is assumed that jobs will “return” from abroad, when in reality they were lost due to technological reasons or will no longer exist in a few years.
…
Since there is no longer any pillar of society left to defend the value of Reason, I propose that we use technological automation to get rid of some of the work done by politicians. We don’t need them. What we need are facts, analyses, scenarios and strategies. What we need is consultation with the people, the study of proposals, the standardization of possible outcomes, the assessment of benefit and cost, the consequences of potential reactions, so that we can make long-term decisions. These sound like jobs for specialists, not for charlatans.
This is where Watson comes in. IBM began to show the result of its work in artificial intelligence when its computer Deep Blue defeated grandmaster Gary Kasparov in a series of chess matches, two decades ago, achieving its first victory in 1997. In 2016, Watson defeated the world champion in Go, the most complex Chinese board game. Watson’s analytical tools are already being used in hospitals to personalize cancer treatments, in school systems to shape individualized education programs for students, and in China where it has undertaken the design of strategies to reduce air pollution.Watson also monitored the 2006 pre-election debates, using its available speech analysis tools, and comparing them with a similar analysis of the videos from the debates between Kennedy and Nixon in the 1960s. What it found was that the civility in the behavior of the candidates has significantly declined since then. At the same time, IBM’s dialogue technologies conducted a massive research effort to reliably identify, collect, and group pro and con positions on a series of issues, such as oil pipelines, tax policy, and the regulation of violent video games. Unlike confrontations that have no relation to actual facts, taking data into account makes democracy more rational rather than an entirely emotional matter.
…
An interesting advocacy, and indeed before Trump begins to unfold his talents! Especially, if one considers, that those who appear as the embodiments of “political power” are, increasingly over the last 30 years, mere showcases, which hide (that is their job) the operation of technobureaucratic mechanisms of each state that are the real “producers” of “political directions” – against which the “politicians” are simply representatives of type! Watson, therefore, may not yet be ready for “central figure of political power”, a mechanic but without passion, a wise and selfless leader. He could, however, be an effective first-line advisor in the exercise of power, and indeed in democracies;
Let no one rush to answer “no”!!! Although simple handheld devices (the remote controls of everyday life: smart phones) do not appear to make decisions on behalf of their users, they appropriately format the “questions” mechanically posed to them, such as a “search” command for example, so as to be “co-shapers” of many “decisions.” The way the gps application works and is used to find a route “from – to” has become so commonplace that no one imagines that a machine takes him by the hand and pulls him with technocratic reliability into the “unknown”… Why would it be so extraordinary for a higher-order algorithmic system (of the Watson type…) to pull many people by the hand, always with technocratic reliability, through the jungle of “pros” and “cons” around an issue, safely leading them to a conclusion that transcends deceit and self-interest?
What Parag Khanna supports (in the continuation of his text) and which all supporters of Watson for president and quite a few others would agree with, is that for (capitalist) democracies to improve significantly, voters should no longer choose persons/representatives. But rather the basic directions of any “government policy”; the type of “problems” that need to be solved prioritarily; the basic specifications that will weigh in the final choice among different answers regarding “equivalent” ones from the viewpoint of data documentation. Once these have been selected by voters, their implementation is a technocratic job (says Khanna); and as such, it can be done to a large extent by artificial intelligence machines…
In other words: from the political use that artificial intelligence could have, the public relations, propagandistic, misleading content of the “work” of political facades; their personification; and all the costs / expenses related to maintaining these parades of rogues called “politicians”, is primitivism. Other parts of it should be abolished, and others mechanized; just as happens in almost all jobs, already from the 3rd and even more so in the evolution of the 4th industrial revolution.
Obviously, among the arguments for such a development would be the claim of technocratic neutrality. If Watson could gain points in his candidacy for the presidential chair, one argument in his favor would be that, as a machine, he would not be biased. However, no machine (and, obviously, no combination of “smart algorithms”) is neutral as technocratic ideology claims. It’s simply difficult to attribute “intention” to the machine; not, however, to its designers…
It doesn’t seem far-fetched, therefore, that at some point—not too distant—the “mechanical resolution of problems” might gain political currency in protocosmic societies… Only that it points to a rather forced possibility. Perhaps, that is, rationality in “democratic decision-making,” apart from a few “major” Watsons on the central political stage, also requires several million “minor” Watsons at the base, standing before the voting systems…
So that the earthly citizens, on the one hand, delegate more and more of their thoughts, decisions, or “solutions to their problems” to machines, while on the other hand they are being dumbed down without even realizing it, that monk of ’48 would be turning in his grave!
Ziggy Stardust
cyborg #12 – 6/2018
- From “Cybernetics and Society” (the human use of human beings), by Norbert Wiener, Papadakis Publishers. ↩︎