Artificial intelligence: the new dependency?

Artificial intelligence: the new dependency?

For example, artificial intelligence is already being used to optimize individual investments in the stock market, or to better organize vehicular traffic by identifying, in real time, the most uncongested routes.

The promotional pitch seeks to sell AI in response to most problems; And without a doubt, many applications can be quite beneficial, on a personal or social level. However, like all technology, the way it is developed responds to specific interests; and currently almost the only entities with the capacity to make the investment and handle the amounts of data required to optimize the systems are large transnational companies: mainly from the United States, but also from China and, to a lesser extent, from some other countries.

The hegemony that these companies have achieved is due, on the one hand, to the key position they occupy in controlling the platforms that connect the different actors, a fact that lends itself to the formation of monopolies. And this in turn allows them to accumulate more data, the main input of this new digital economy. Then, and especially when it comes to transferring public services or critical functions to AI systems managed by these companies, a contradiction arises between the goal of maximum profit for the company and the demands of the public interest.

One of the most obvious risks is an eventual failure or hack in a vital system (such as the electricity grid) or a high danger system (such as self-driving vehicles). Possibility that increases if the responsible company tries to increase its profit by reducing spending on security.

But serious implications and challenges arise in many other aspects, particularly with regard to human rights or legal gray areas; as well as in matters of sovereignty.

In developed countries (particularly Europe), the debate on the implications of artificial intelligence is open and the development of frameworks of principles and rights has begun, which consider issues such as:

- Robots and AI systems programmed to make certain decisions sometimes have complex algorithms that it is impossible to know exactly how and why they made that decision and not another. So who is responsible for the consequences of these decisions?

- Who owns the data that computer systems collect from sensors (for example, from a city) or from users (with or without their consent or knowledge)? What implications would it have for who (s) benefit from the economic returns they produce?

- How to prevent intelligent systems from deepening exclusions and discrimination (intentionally or not)? In fact, there are already many cases where it is evident that social prejudices are reflected in the same algorithms.

Possibly one of the most acute problems would be the impact on employment due to robotization or automation of the production of goods or services. There are forecasts that employment in many sectors will disappear, and that new jobs will be insufficient to absorb all the displaced people; Among the most vulnerable sectors are professional drivers or sales personnel in supermarkets and warehouses. For this reason, there is increasing support, in developed countries, even among the business sector, for the idea that it will be necessary to establish a universal basic income for the population that is left without paid employment, which would be subsidized through income transfer policies of the ultra-profitable companies in the AI ​​sector.

Every time, other analysts consider that the danger of job loss is exaggerated, at least in the short term, (perhaps for political reasons: a worker with fear of losing his job will be more docile), since if it were true that robots are massively replacing workers, there would be strong growth in productivity, which, at least in the case of the US, is not recorded. [1] Average growth is only 1.2% per year in the last decade and only 0.6% in the last five years.

But there is no doubt that there is a transfer of wealth to companies that concentrate power in the AI ​​sector (sometimes known as GAFA –Google, Apple, Facebook, Amazon–, or GAFA-A, including the Chinese company Alibaba); enrichment based on data accumulation and processing,

The impact in the South

In Latin America, so far, there is little debate on these issues. However, we can estimate that the impacts will be significant and in the relatively short term. On the one hand, the changes in the North will undoubtedly have consequences in the South. For example, as robotization and automation progress, certain production lines that were moved to southern countries to benefit from cheap labor would return to the North. In fact, it is already happening: in India, for example, jobs in the information technology sector, in particular call centers, have been sharply reduced. On the other hand, the contracting in the South of AI systems from North suppliers, for example to improve public services, will mean new forms of extraction of wealth and data and therefore new forms of dependency, greater gaps between North and South, etc. It would be important to carry out studies that measure the real repercussions in our countries and to estimate the potential impact.

In an op-ed recently published in the New York Times [2], Kai-Fu Lee, (who heads a Chinese venture capital firm and chairs its Artificial Intelligence Institute), presents the outlook in rather stark terms: for For the foreseeable future, while AI is a long way from being able to compete with human intelligence, he recognizes that it has the ability to reconfigure the meaning of work and wealth creation, which will trigger the wide-scale elimination of jobs, leading to to unprecedented economic inequalities. For this reason, he considers it inevitable to introduce income transfer policies from AI companies with high profitability to sectors without employment, which will be feasible - he says - in countries such as the US or China, which have the potential to dominate the sector. But, since AI is an industry where strength begets strength, most countries will be left out of that possibility, so they “face two insurmountable problems. First, most of the money that artificial intelligence produces will go to the United States and China. " And second, having growing populations will become a disadvantage, due to the scarcity of jobs.

So, he asks what options will be left for most countries that will not be able to tax ultra-profitable AI companies: “I can only predict one: unless they want to plunge their people into poverty, they will be forced to negotiate with the country. to provide them with as much artificial intelligence software - China or the United States - to be essentially economically dependent on that country and accept welfare subsidies in exchange for the 'mother' nation's AI companies to continue obtaining profits from users of the dependent country. " The author estimates that US companies will dominate in developed and some developing countries, and Chinese companies in most developing countries, an economic arrangement that "would transform geopolitical alliances."

It is certainly a forecast influenced by the Chinese geopolitical perspective, but we highlight it here because it is rare that the business sector wants to acknowledge this reality. You can think that there would be other ways out; However, with the current inertia in most Southern countries in the face of this still poorly understood reality, a scenario similar to that envisaged by Kai-Fu Lee seems quite likely. The South would remain in its role of supplier of food and raw materials and its dependence on the North would deepen.

There is not much time to react, as highlighted in his recent visit to Ecuador, by the former Greek finance minister, Yanis Varoufakis, who warned that the current economic model of that South American country will barely last five more years and then - if there is no technological replacement - it will be left out of the value creation chain. "Technological change is moving rapidly against primary producers: low- and middle-income countries that depend on physical trade." While praising the sophistication of Ecuadorian financial policy in the face of dollarization and foreign debt and for the redistribution of income, he considered that the current challenge is to find a similar sophistication in the technology sector, emulating, for example, Estonia or Iceland, with a policy of technological sovereignty, so that it becomes an example for the region and for the regional integration process.

Meanwhile, the transnationals in the sector are rushing to break down any remaining barriers to their global dominance over markets and data. They advanced their agenda, with very little resistance, in the e-commerce chapters of the TPP (Trans-Pacific Treaty - now defunct) and TISA (Trade in Services Agreement - frozen for now); So the bet now is to open negotiations on "electronic commerce" at the World Trade Organization (WTO) [3].

Undoubtedly, the challenge of the new digital economy calls for a clear and forceful political will, but also to seek alliances. Due to the size of the investments it requires, it is unlikely that any Latin American country on its own can find an adequate outlet; But a bloc of countries - such as UNASUR - would have a greater capacity to develop response levels, at least to assert regional sovereignty in some critical areas. It would also allow it to accumulate more bargaining power vis-à-vis the AI ​​powers and their companies, such as in global instances where governance policies are defined.

By Sally Burch
British-Ecuadorian journalist, executive director of ALAI.

[1] See Dean Baker, "The Data Defying Job-Killing Robot Myth", here
[2] Kai-Fu Lee, "The Real Threat of Artificial Intelligence", New York Times, June 27, 2017. here
[3] See Sally Burch, “The e-commerce agenda at the WTO”, here
Article published in the magazine América Latina en Movimiento: Faced with challenging scenarios 07/03/2017, see here

Video: Conceptual Dependency in hindi. Artificial Intelligence. #27 (July 2021).