Artificial Intelligence and Taxation


Tulio Rosembuj. Septiembre de 2018

The algorithm is not a mere mathematical formulation. It is a programming design based on the data that supports it, the more significant in better quantity and, also, the source of rules of conduct based on predictions and probabilities of data on human behavior, to modify it, channel it or direct it. The algorithm allows us to transform our subjectivity as a person into code; but, in parts, in fractals, turning the individual into dividual.

The mathematical model is threatening because it is a weapon loaded with prejudices, weaknesses, errors. The past guides the compulsive programmers towards their projection in the future. Replica of any dictatorship with pretensions of intellectual superiority. Artificial intelligence constructs the dogma of technological supremacy on sandy bases because it attemps to control, vigilance, punishment as means for profit or security or defense.

The social power of the algorithm is nourished by its public relevance, its institutionalization, its necessary subordination of the public and social sphere to its framework. Algorithmic authority is the form of influence that is exercised on the behavior of individuals, with no other pretension than its program, design, and practice. So, we are talking about a source of de facto, opaque and secret duties, rights and obligations, a cause that adresses by itself the inclusion or exclusion of its consequences, the retribution or punishment on people as programmed by other people.

The social power of the algorithm lies in the ability of networked computers, in the cloud, but, above all, in the collection and capture of personal data. Personal data is the raw material of digital wealth. There is no more valuable currency in the millennium than the personal data and gratuity also makes it the source par excellence of the behavioral surplus (behavior surplus) of large organizations involved in the Internet. The function of data is the construction of the digital person, whose identity is partial, dividual, because it is based on a generalization, inference, correlation, directed to the investigation or control of the behavior of the categories, segments, groups, social to which it tries to influence, motivate or persuade.

Free appropriation the data is the starting point for the production of predictive products based on the manipulation of the behavior of consumers and citizens. The data ceases to be of the person and becomes the ownership and possessory secret of the organizations, which transform them in the profiling of individuals that are invisible and ignored to the users. The surveillance economy is the economy based on the control and monitoring of the person from their data, which can be used by anyone, in any space, at any time and with an eternal, reiterative, reuse, with a clear motivation for profit or information.

The social value of privacy is a value that is not individual but is of interest to all society. The General Regulation of Data Protection (GDPR) of the European Union adopts the superior principle that the protection of Individuals concerning the processing of their data is a fundamental right. The scope of the European standard makes clear its emphasis on the respect for human dignity as the basis of data protection that cannot be fully transferred. Our data is part of each of us, fractal segments of our identity and identification and this by human nature. The separation between the person and its data is irremediably artificial.

The European Union provides an innovative regulatory framework for the protection of privacy in the face of the emergence of data surveillance, where public intervention as a last option point to the responsibility of economic agents and an attempt to an approximation to the algorithmic mystery. It is clear that there is underlying reasoning, that the design, practice, complexity of the algorithms is potentially harmful and, above all, unknown in their projection on the person.

Machine learning is the algorithm raised to the cube. The program aspires to learn by itself, from the accumulated data and through generalizations that start from the own data and the instructions of its design. It is a vehicle of predictions, associations, inferences, everything but information and knowledge obtained from the scientific method. Generalization does not distinguish between correlation and causality, probability and truth.

The algorithm that learns by itself does not deduce the glorification of the program, design, and application; but, it is triggering the risk of being at the mercy of the subjective, capricious, partial decision of the programmer. The mystery is the abstract model of data that no one understands or can explain and that determines our behavior without our knowing it. Algorithmic power is baptized as a mathematical weapon of destruction.

Robotics, whether physical or virtual, authorize to consider it as a risky application of the machine on people. The enhancement of the robot be it more or less autonomous or not; be it for the factory, the home or the war; is the worst threat that artificial intelligence carries. It is an artifact aimed at displacing the person from his family environment, from work and transferring death from a distance and far away. The robotics cannot be the way to avoid responsibility and obligations of the programmer, developer, designer, trader or manufacturer.

Automation is the name of artificial intelligence, a mixture of digital good and vigilance. The hurricane of creative destruction of technological innovation has a vocation for unlimited power, deregulated and detaxed, where the rules are imposed by the agents who carry it out.

Automation without a public or social intervention is a conduit of market manipulation and disinformation and displacement of work by the machine and the algorithm. Automation should be subject to values, norms of law that reduce their aspiration to control and monitor the behavior of the person, the citizen, the consumer and without assuming the negative costs, the externalities, of their activity. This means taking evidence of the free use of private information; of the reversion of the public domain of that fruit of the appropriation and the cartography of the predictive products, manufacturing of the information apprehended and reworked.

There is no technological future focused on automation without the person at work, working negative externalities are a consequence of the lack of responsibility of the economic agents that are beneficiaries of artificial intelligence and on them should rely the effort of solidarity and the common good. We are at the doors of systemic risks for lack of precaution and lack of control of hazards, precisely as it happened with the financial system in 2008. The contamination of data and negative externalities illustrate the public and social damage of artificial intelligence left to their free will. The collaborative economy (gig economy) serves as an example to highlight how innovation disguises the abuse of labor exploitation as if it were something different.

Korinek- Ding Xuan Ng refers to digital innovation as a superstar sector in the economy, “in which a small number of entrepreneurs or professionals distribute their output widely to the rest of the economy.” This occurs because information technology requires fixed costs but can be reproduced infinitely at zero marginal cost and, consequently, geometrically increases the benefits to scale. The digital innovation of the surveillance economy also provides a monopoly and monopsonic power to the innovators in the market without cost. Predictive products are generated at zero cost (The 2017 Macroeconomics of Superstars Anton Korinek Johns Hopkins and NBER Ding Xuan Ng Johns Hopkins November 2017-stats-forum / session-3-Korinek.ashx)

The authors highlight the effects of digital innovation on the cost savings it produces to its producers and the market power they achieve. But, assuming that a work shift is verified that is now redundant and causes the correlative decline in the demand for work and, on the other hand, the origin of a monopoly income that they call “the superstar product share”, based on their power over the market.

“We argue that this represents one of the fundamental driving forces behind the rise in inequality in decades. …Although monopoly rents for superstars support their investment in digital technologies, the overall level of such rents is socially excessive.”

Technological innovation is a source of inequality and excessive monopolistic income. The surveillance economy makes this inequality and excess profit permanent. The behavioral surplus improves the first advantage of the digital economy thanks to the manufacture of predictive products.

Here are no reasonable doubts that prevent a dark perspective of the immediate future of human capital. And it does not help to settle for training, education, the emergence of new professions, because, in any case, it would be about far-reaching changes to alleviate immediate or very serious misadventures.

The first step of the tax system is the equality of qualitative and quantitative treatment between labor income and capital income. Or, in other words, capital must contribute equally to or more than the income from work.

The second is eliminate tax benefits-exemptions, rebates, reductions-to those who no longer need it.

The third is to extend the tax base of Corporate Tax, eliminating all deductible expenses that are not justified in the face of the investment decision and the accelerated depreciation of capital costs, aimed at taxing the gross income of the company.

The fourth is the introduction of a Tax on Autonomous Work.

Finally, the new taxation on artificial intelligence as a whole should support a VAT income type tax on the organization, a tax on the surplus measured on sales or on the excess profits; a tax on personal data that limits the market manipulation and the negative externalities of the algorithms, something like a Tobin Tax or a Bit Tax or the European Digital Services Tax, and, finally, the transformation of the data user into a data worker.

The sources of the surplus that must be taxed are the exceptional yields of the predictive products, of the robots and of the propaganda of precision. And the taxpayers are all those economic agents, digital or not, that take advantage of the automation. The tax must not target the robots, but one of its outstanding elements is the fight against unemployment. The social costs of the robot are costs that must be borne by the organizations that produce and market it. We speak of a tool that increases the income of the activity of the producers and the behavior surplus of the organization.

Mass unemployment is the red light of robotics, the risk to be avoided; market manipulation and misinformation is the so-called propaganda of precision. In both cases, we face the negative externalities of automation. The tax challenges arising from the digital economy come from two fronts: precision propaganda   and robotization. In summary, the surplus comes from the manipulation and misinformation of personal data and the mass expulsion of human work through robotics.

The automation tax must necessarily subject both activities to the contribution to public expenditure. The digital manipulation of the market consists of the systematic application of techniques of motivational behavior, persuasion and influence on consumers and misinformation of the citizen in the orientation of their decisions on public policy. Precision propaganda does not distinguish between market and politics. It deals with the capture of personal data and the development of predictive products to achieve the benefit or control through data surveillance. Profiling abuses the ignorance of the individual because its purposes are known by those who design and perform it, but not by those who suffer it.

The Equalisation Tax is one of the first responses to the generation of a behavioral surplus of precision propaganda. The tax, experienced in India, has been improved by the Union European Union through the proposed creation of the Digital Services Tax.  It is a tax that aims at the creation of value by the user: the monetization of the contribution of the user. In other words, by the value obtained by the managers of the economy monitoring the personal data of the users.

The object of the ISD is the accumulation and regular monitoring of personal data derived from digitization, although it is concentrated in precision propaganda and companies with a high digital significance. It is an essential contribution because the behavioral surplus of the company is located where the economic value is created and the user’s contribution to its origin. The EU proposal overcomes the OECD digressions and broadens the foundation evidenced by the Equalization Tax in India.

The European Union assumes the following right criterion:  digital economy is based on new, original intangible assets, creating wealth, such as use of data and predictive methods to extract value from user data. Digital economy is first and foremost an economy of vigilance. It forces the development of current concepts of the permanent establishment; the idea of ​​a Common and Consolidated Base of the Corporate Tax in the EU (formulary apportionment) and the search of solutions of consensus at a global level, in particular in the G20.

It is useless to refute that one of the serious failures of the OECD in its proposals on digital economy lies in the insufficiency of its criteria to overcome the traditional concept of permanent establishment, reduced to mere physical or material presence, which prevents the subjection to the tax of the digital economic activity where the value is created. There was no decision on the virtual permanent establishment as a sufficient condition for taxation in the territory where it obtains its benefit, even if it does so without physical or material presence.

The European Union takes a step forward: the permanent establishment must assume the changes produced by the automation enabling each State to tax the economic presence of the non-resident in its territory, although it does not have a material presence and, at the same time, update (if not profoundly change) the transfer price rules and attribution of benefits “to tax the benefits where the value has been created without a physical presence.

The Consolidated Common Corporate Tax Base project, if approved, would provide, in turn, with a formula for sharing global benefits of a multinational company among the member countries, where it acts, in proportion to the sales in each one of them, which exceeds at arm’s length and allows the verification of transfer prices. Finally, the EU solution must be extended to relations with third States, which implies the adaptation of the treaties to avoid double taxation following the defined lines of the Single Digital Market. This requires a common positioning of the Member States, to prevent disaccord in the European and international tax system.

The debate has just begun. At the center of artificial intelligence is a huge improvement in the interconnection between people and traffic of information that the Internet, in various ways, makes possible. Contemporaneously, the perspective is obscured by the catastrophic prophecies that may occur in people’s daily lives. Technological innovation, as in the eighteenth and nineteenth centuries of the industrial revolution, is accompanied by good news and bad news; of desirable changes and unwanted changes.

Polanyi writes: “At no other time did liberal philosophy fail so clearly in its understanding of the problem of change. Acuciated by an emotional faith in spontaneity, the attitude of common sense regarding changes was abandoned, in favor of a mystical acceleration to accept the social consequences of economic improvement, whatever those might be. ”

Now we are immersed in significantly similar crossroads. The unconditional acceptance of technological change is imbued with spontaneous mysticism that proclaims its supremacy in the face of social, economic and political obstacles, which it contributes to creating and whose solution it leaves in free will, sacrificed for the sake of speeding up the process at any cost, including, the sacrifice of the common good, the minimum social needs, the distribution of benefits and sacrifices among all.

And, again Polanyi teaches the devaluation of the welfare of the community during the industrial revolution because of “the corrosion of a brutal utilitarianism combined with uncritical confidence in the presumed autotherapeutic virtues of unconscious growth”.

Artificial intelligence is a true example of the maximization of short-term benefit based on surveillance under the mysticism of technological inevitability. Nothing that opposes the interconnectivity between the algorithm and (the data) of the person is right. Everything that opposes it is terrible.

Liberal philosophy begins as the revolution of the rich against the poor who are expelled from their uses and natural habits, are cornered in poverty and misery, children are exploited, and women and factories become monuments of the exploitation. The divinity of progress is ineluctable when it requests the offering of danger, abandonment or destruction of the human being.

Industrial revolution relied on a self-regulated market, without law or regulations, that would model the safeguarding of the person and the community. Now, artificial intelligence campaigns for its privileges, without statute nor norms that guide it, limit it or force it. The consequences are apparent, the negative externalities do not cease to be denounced and, in the absence of an active and interventionist public sphere, the autocracy of the algorithm spreads, as a de facto source of the own law. The social power of the mathematical instrument as a means of surveillance aimed at a profit or social or warlike control (killer robots).

Hence, the urgency of public norms, of imposition, which must seek to minimize the negative externalities of artificial intelligence, even at the price of delaying its rapid passage. The pause introduces an element of critical reflection when uncertainty dominates. Caution is not a luxury; it is a demand before the ignorance of the unknown. The accelerated pace of change can be cushioned without changing technological innovation, saving social morality, values ​​and principles of institutional coexistence.

Artificial intelligence would not be what it is without the Internet, the power of computers and the interconnection in the network. It is not by chance that its inspiration is before the Internet or that robotics was a cult of science fiction. It would not have been possible without the explosion of the algorithm, cryptography, symmetry to asymmetry and electronic transmission of bits and bytes, digital and virtual goods between computers.

The Algorithmic Society in the definition of J.M. Balkin is “a society organized around social and economic decision making made by algorithms, robots and artificial intelligence agents; that they not only make decisions but, in some cases, they also carry them out. ”

Algorithmic decision making is not neutral. “Because human beings program predictive algorithms, their biases and values ​​are embedded in the software instructions.” (D. Keats Citron & Frank Pasquale, The Scored Society: Due Process for Automated Predictions, Washington Law Review, 2014, 89: 1.)

Personal data is the fuel that feeds the machines, instruments, and artifacts used by artificial intelligence. The question lies in the fact that the algorithm obeys to the governance of humans over other humans, based on the data collected and treated.

Algorithms are human creations, and there is a definite human responsibility for their use to influence, motivate, persuade to change the behavior of other people, either through precision propaganda or through machines, instruments, artifacts, aimed at improving their perspectives of daily life, for example, your job.

The algorithm is a program of an author, invention or creation, which never detaches itself from its work, as much in its successes as in its errors. Copyright is correlatively the responsibility in case of damages or risks, on third parties. It is not a book, but a weapon, an instrument loaded with promises, that can destroy or diminish individuals.

The algorithm is somewhat similar to financial derivatives before and during the financial crisis of 2008. At that time, Bernard Buffet warned against its potential destructive power, which was later confirmed by the facts. Now, Cathy O`Neill resorts to the expression of weapons of mathematical destruction. (Cathy O’Neil, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy, Crown, 2016.9

Again, the problem is not the algorithm itself, after all, a mathematical formula, but the human creation and its use as a weapon of mass destruction, supported by the concentration of computing power, massive data collection, of elaboration of predictive products.

The dark side of artificial intelligence is dangerous because it involves damage or threats to the life of each one at critical moments, e.g., being admitted to one school versus another; being approved when applying for a loan; searching or staying at a workplace. Data surveillance reveals itself as a hidden, secret model of reward or punishment.

Deregulation is the realm of the algorithm. If this situation is not modified in a different direction where transparency, responsibility and equity prevail, a scenario of maximum vulnerability and perpetuation of social inequality is generated. The exact reproduction of the past with the intention of the future.

The powerful attraction of the Internet can not subvert its route, converting the desired interaction between people in their sin, just because the excess of benefits and the ambition of control dominate the oligopolistic information market. During the financial crisis of 2008, somebody accused the holders of subprime mortgages of being responsible for the hiring of such toxic products, yielding to temptation, which was subject to persistent commercialization by financial entities.

Regarding digital failures, the user is being blamed to apply, click and connect even if that is systematically modeled by large organizations to exploit their interests, anxieties, anguish.

“The ease and speed with which internet users trust tech companies with the data they are wildly out of proportion with the risks they are assuming in doing so, in no small part because … the companies could not anticipate the risks themselves.” (Jh Herrman, Holding Facebook to account, The New York Times International, April 17, 2018.)

Why organizations can not anticipate the consequences of their actions? It’s not only negligence or greed, but also because digital arrogance does not allow them to notice, as in the matter of climate change, that their disordered activity carries negative externalities, systemic risks: ignorance prevails over their predictions and probabilities. Failure of responsibility is closer when least examined the own mistakes of conduct, which does not serve as an excuse.

The digital market is reflected once mathematics materialize as the next generation technological source in applications, instruments, devices, machines. Even so, nothing would have been transformative, if all this is not based on personal data for predictive production.

The logic of technological capitalism is based on the mathematical machine in a broad sense aimed at the mass production of informational goods that are nourished free of the collection of personal data, infinite, global, eternal first subjects, without whose power the machine would not serve for the intended commercial and control purposes.

The surveillance economy is the algorithm plus personal data. The excess of profits is the result of a market free of public interference and whose prices are self-regulated by the same agents that carry it out. The consequences of the digital market, without rules, are established by the power of platforms and social networks on people and society as a whole.

The absence of a law opposed to the free activity of data surveillance admits two reservations: that the new law is a consequence of self-regulation sought by economic agents or worse, that institutional corruption diverts the public sphere in favor of private interests. In both situations, the general, collective interest suffers.

The digital market becomes the point of arrival of the entire economy because that is where the highest profits are obtained. The difference between digital and traditional economy is irrelevant because if it does not access the first, it will be out of the market, including its own. Either the adjustment to the algorithm is verified or it will disappear. The traditional economy said as the one based on the production of goods, services, materials or physics succumbs to the behavioral surplus of predictive products, informational merchandise, including robotics. Digitization is the only way that can allow its survival because it joins the average benefit of the traditional product – the car, the refrigerator, the bricks – with the intangible component of the free and inexhaustible raw material of personal data, as a source of surplus.

The ownership and structure of digital technology is a development of the appropriation of personal data of individuals and their conversion into intellectual property. The confiscation of data is the premise of predictive products, of profiling, trapped in the domain of each and added to a category, segment, social group, to define, motivate, influence or change their behavior. Free appropriation or capture of personal data is the foundation of the digital explosion.

“Why complex machines are costly are not profitable unless large quantities of goods are produced. They can only work without loss if the production of the goods is reasonably assured and if the production should not be interrupted by the lack of the first materials necessary to feed the machines. “(Polanyi).

Mathematics and cryptography are not worth a machine, without the fuel of personal data that provides the typical predictive products of the surveillance economy. The spectrum is broad, comprising both the predictive intangibles and the Internet cryptocurrency, such as Bitcoin and others.

J.E. Cohen, adapts the Polanyi scheme, focusing the surveillance economy within the framework of three criteria: the ownership of intangible resources; the datification of the basic factors of industrial production and the absorption and reworking of profiles within information platforms. The harvest of data recalls the appropriation of raw materials, wheat, oil as if it were a public domain. The possession of the data becomes the source of secret ownership of digital companies and hence the behavioral excess of benefits.

The effects of innovation do not strengthen a hypothetical free market of ideas, but a service of extracting benefits from an informative environment, under the cover of processes of truth and discovery. “The result has been a series of object lessons in the law of unintended consequences It is high time we faced up to the fact that we have the technologies of freedom that we saw we want. “(Julie, E. Cohen, Technology, Political Economy, and the Role (s) of Law, Law and Political Economy, June 8, 2018).

The legal framework can not delay in channeling the pace of technological innovation in a direction compatible with the preservation of the human dignity of the individual, the conservation of human work at work, the protection of market manipulation and misinformation.

It is true that it invades a certain uneasiness when the State is aware of the unconditional subordination of the free political function to the company and the market over the information economy. Institutional corruption (L. Lessig) permeates the neutral and equitable public intervention with doubts. However, if it is not the public sphere, there is no way to manage digital algocracy. The State is the last edge of democratic governance and if that is not the case, let’s prepare ourselves for the oligarchy of the algorithm.

The consequences are not hard to imagine. The propaganda of precision is equivalent to a new system of information and disinformation, which affects, at the same time, the truth of the facts between individuals and the public sphere itself; the subordination of the collection and processing of personal data to private investors; the monetization of the data, voluntarily shared, for the immediate benefit of the organizations that produce the predictive products or their use as privileged means of social control in the hands of agencies, entities, with unknown purposes. Robotics, as a perverse culmination of social injustice, unless its effects are fixed to prevent the not so obvious superiority of the machine over the worker.

The predistributive legislation covers the entire legal system.

First, the reform of intellectual property, patents, copyrights, trademarks, know-how, because, the first wealth is personal data, free and extracommercium, which, in theory, cannot be subject to another property than that of its owner. If there is a peremptory sphere of intervention, it is intellectual property, because it hurts the common thing, the public domain, in a direct and evident way. The transformation of the public good into a private right of exclusion must be remedied by opening ownership to regimes that are more open and participatory because the user’s work in creation, invention, knowledge is irreplaceable, even if it is not legally recognized.

Second, the presence of few companies in the appropriation and application of the surveillance economy proposes to face its dominant oligopoly character. The reform of the vertical and horizontal concentration of platforms and social networks should moderate their current oligopoly and monopsonic nature.

Third, the change of labor law suffers from the intensity in the protection of the value of human work. It is about preventing their replacement by the machine and ensuring the survival, in any case, of the person who is a victim of the process of technological change. And this requires placing unemployment as a priority, until the definition of new jobs, new professions, skills and knowledge that current workers lack and will need time to assimilate is achieved.

The substantial reform, however, is the study of Data as Work, a suggestive hypothesis through which the user is qualified as a data worker in the co-creation of digital value. This implies an orientation towards new categories of workers, their organization’s unions, the classification of jobs, their remuneration, the rights and guarantees they deserve.

Finally, although not, ultimately, the intervention of the Public Administration in full “command and control”, designing a penetrating scheme of authorization of algorithms, as if they were pharmaceutical chemical drugs, susceptible to be carriers of arbitrary, discriminatory decisions, capricious or unemployment. The algorithm is the drug of the millennium and mathematics cannot be disassociated from monsters of reason, as it was, for example, thalidomide, and this means establishing test procedures and preliminary tests, knowing about the content of their decisions, studying their purposes and once approved, evaluate their consequences in the application.

The virtual currency, diverse but concurrent with the digital economy, also requires a particular political approximation of the cryptographic formulas used, which draw escape routes from the algorithms to any mode of traceability between transactions and people. The marriage between the algorithm and the cryptography threatens a black hole in which there will be no borders for tax evasion and money laundering. The new currency of the Internet, without prejudice to its recognition as currency, need a global and fiscal legal framework that prevents the unregulated overflow from the side of concealment and maximum secrecy.

Tax also has its role. On the one hand, the digital company is virtually immune to general taxation; on the other side, artificial intelligence, automation, in itself, as an exceptional source of benefits cannot be left out of the contribution to public expenditure. Who will pay the social costs of what will come?

Artificial intelligence, automation, can be taxed in four ways: (i) a VAT income type tax; under the subtraction method, not deduction as the Community VAT, (ii) a tax on the surplus measured on sales or on excess profits; (iii) a charge on the collection and treatment of data and (iv) the consideration of data as work and the right to remuneration of workers, who are not consumers, of data. They would be the ways of internalizing the negative externalities that they produce in the economic and social system.

The robot is an artifact that contributes to the benefit of its producer, and that is therefore responsible for its introduction into the market whenever it raises negative externalities. It does not seem defensible an autonomous tax, but its qualification and quantification as a machine in the Corporation Tax, or in some new initiative aimed at artificial intelligence as a whole.

In parallel, the expansion of the tax base of Corporate Tax should be considered, eliminating the qualitative and quantitative discrimination of capital income in favor of equal or better treatment of labor income.

The challenges of the digital economy to the tax system are in their infancy. Now, the first steps begin. The European Union assumes leadership in the face of market manipulation and misinformation through the proposal of a Tax on Digital Services, which improves the Equalization Tax of India; the adoption of the Common and Consolidated Tax Base of Companies and the permanent virtual establishment.

The position of the European Union raises sharp contrary reactions, for the most part, from the USA and the OECD. It can not be said that these are solid arguments, but rather an ideological campaign in favor of the inevitability of the advantages of digital technological change, despite the social disruptions it causes.

Thus, M. Herzfeld argues that European policy has more to do with concerns about the structural economic changes of digital innovations than with taxation. In truth, his article is quite a proclamation: “Taxation of the Digital Economy: A false fiscal policy”. The reasoning is childish because if something cannot be predicated is the divorce between technology and the market, the State and citizens. What seems undeniable is that the social costs of digital technology cannot be discharged over those that do not. The tax is part of the solution, not the problem, whatever the answers that are chosen. Presumably, what is not wanted, in good faith, is the repetition of social injustices that accompanied the industrial revolutions of the eighteenth and nineteenth centuries.

The fiscal issue is not a mere escape, but a necessity to provide the minimum equality and justice to the economic dispossession of the most disadvantaged. The author is right when she argues that digital taxation has less to do with the null tax or almost paid by technology companies, American, than with its high profitability. Indeed, the issue is the oligopoly of Apple, Oracle, Cisco, Google, Microsoft, Facebook, Amazon, Netflix, among others.

That market dominated by the American companies does not request to declare their fiscal immunity (or does it?) Allowing them to persevere the negative externalities of their economic activity. The tax appears as one of the few instruments available to the States to repair the damage, the damage, the irresponsibility of technological disruption.

M.Herzfeld also interprets that the BICC of the Corporate Tax  is a pretext to “modify the sovereignty of the States (of the EU) to establish their own fiscal policy on the societies.”, Unsustainable argument because it is an european unión  project that precedes long the evolution of the digital economy and, unlike what it says, can contribute to building a stronger fiscal integration than the current between the States, to begin with, the elimination of transfer prices and the sharing of tax revenues between all according to the expected distribution formula. (M. Herzfeld, Digital Economy Taxation: Fake Tax Policy, Tax Notes International May 7, 2018).

The Equalization Tax is another controversial issue. The position of the EU is more determined, for example than that assumed by the OECD. The first difficulty consists in its legal nature. It is not a tax on the sales, so it can not be said that it is against the EU VAT regulations. It is a direct tax on the intangible income of the organization included in the provision of digital services. It is true that its approval requires the unanimous vote of the Member States; but, it is not less than nine would be sufficient to promote the mechanism of improved cooperation ( TFEU).

The virtual permanent establishment configuration also arouses critical criticism. For some, the idea of ​​anchoring the application of the tax to the creation of value through the user’s location factor within a given territory is unacceptable. Even at the price of accepting that physical or material presence does not ensure its imposition in the digital economy. The benefits should be taxed in the State in which there is a significant digital presence, and its distribution with other States is based on the number of users and the volume of information collected in each particular State. “The merit of the proposal is that it exists and is, therefore, a step forward in the development of ideas about how to tax digital services.” (F. Vanistendael, The Level Playing Field in Digital Taxation, Tax Notes International May 14, 2018).

Digital technology cannot be separated from politics and the market. The tax would be the main beam for the oligopolistic power of organizations, platforms and social networks to contribute to public spending based on their ability to pay. But not only. All added economic activity that brings digital advantages must be proportionally taxed.

The automation traces fingerprints, hardly visible, and physical traces, the artifacts that follow, that can be recognized. Precision propaganda and robotics are the sources of the excess benefits of surveillance economics. But, also, there are prominent examples of obfuscation. The sharing economy cannot be the outgrowth of the loss of social orientation unless the worker is confirmed, working, the professional, serving; the entrepreneur, producing and the intermediary entity, mediating (Uber, Airbnb, etc.)

The technological loyalty to the company and the market leads to desolation. It would be possible to establish a countervailing balance between advantages and disadvantages that lessen the effects of digitization. But this requires a State and society – or several States and communities – to draw lessons from history.

Artificial intelligence, in its referential cult, encourages processes of poverty and misery similar to the previous industrial revolutions, which will come from the substitution of the person in work for the machine (robots) and the manipulation and misinformation of the consumer and citizen through the predictions, generalizations, inferences, algorithmic correlations.

This industrial revolution will not speak of “dark Satanic Mills” (W. Blake); but of “dark Satanic Algorithms.”