The introduction of Digital Tribulations, a series of intellectual interviews on the developments of digital sovereignty in Latin America, can be read here and the entire serie can be found here.
When I arrived in Argentina in March 2026, my friend Simone Robutti, a political organiser with Tech Workers Coalition based in Berlin, had put me in contact with Esteban from AGC (Asociación Gremial de Computación). A gremio is the specific Argentine term for a trade union, which, besides negotiating wages, does all sorts of activities: managing their members’ healthcare, owning recreational facilities, running vocational training centres, and more. They function, in practice, as a parallel welfare state organised by sector. With the nice weather of the end of the summer, I walked into the union building in the centre of Buenos Aires. As I went upstairs, I was welcomed with coffee, and we sat around the table in the AGC’s meeting room. 
As Esteban self-identifies as a Peronist, some background on this very argentinian political movement is necessary. A few weeks after the interview, I attended the march in Rosario marking the fiftieth anniversary of the end of the military dictatorship. There, another proud peronist, profession psychoanalyst, objected to my doubts about peronism, which to my eyes is not really a leftits movement, by saying that I was using “Eurocentric categories”. Peronism, he said, was neither left nor right: those are the two wings of the peronist bird. The description nicely reflects how Peronism officially presents itself, and it captures something real. the movement that emerged around Juan Domingo Perón in 1945 was not a class party but a coalition that included a bit of everything: organised labour, nationalist industrialists, the Catholic Church, and sectors of the military. 
Over the years, Peronism delivered concrete results, like wage increases, paid holidays, women’s suffrage, a social infrastructure of healthcare and workers’ services that the union system still administers today, as Esteban describes. But the same ideological flexibility that made Peronism a broad popular movement also made it incapable of resolving Argentina’s core economic problem: a country whose wealth is concentrated in agricultural land that was never redistributed, and whose governments have oscillated for eighty years between taxing that wealth and backing down when the landowners pushed back. Back in Italy, I talked to a friend who cited the case of Benetton: the italian clothing company bought roughly 900,000 hectares in Patagonia, making it of the largest private landowners in the world (with obvious contrast with indigenous communities like the Mapuche). 
The same pattern repeated with finances. A supposedly peronist like Carlos Menem (1989–1999) run a radical neoliberal policy, privatised state assets and pegged the peso to the dollar with disastrous consequences; later, Kirchnerism redistributed during the commodity boom while hiding inflation figures, and each cycle ended in external debt, inflation and a political class that had deferred the structural choices one more time. Milei’s rise is the accumulated consequence of those deferrals; funnily enough, with all the people I talked to in over a month, only one veterinary admitted that had voted for Milei “as a protest act”. Like in Italy with Berlusconi, people were too scared to admit it. 
With this, I am not saying that Peronism is not politically viable or that it should be abandoned; however, I see the need for a clearer political stance between the right and the left: as Bobbio knew, there is no escape to the political compass. 
As of today, the AGC is torn between Milei’s disastrous policies on the one side, and the very concrete threat of artificial intelligence on the other. The interview with Esteban – who was always very kind and who put me in contact with many others of Argentina’s tech worker – covers the AGC’s legal history within the Argentine union model, the obstacles created by the current government’s hostility to collective bargaining, the debate within the sector about AI and layoffs, and the union’s work through its Observatory and Laboratory on wages, tax policy, and digital sovereignty.

***
What is your background and what brought you to this union?
I can tell you the professional side: I have a degree in Literature. I am also a university-trained political analyst, that is, the intermediate degree from the Government and International Relations programme. And I am also a student of the Bachelor’s in Mathematical Sciences. Due to a natural political inclination, I have been politically active in Peronism for many years. And it is because of that activism that trade unions interest me. Also for academic reasons — being a Mathematics student — the area of computing caught my attention. In any case, it was partly a coincidence: I met someone and ended up here, at the union. It was a political matter, driven by the need for an IT union in Argentina. Out of a concern for social justice, and also because of a political reading of the big tech companies, their problems and the challenges they bring, it became clear that such a union simply did not exist.
And what is the history of this union? How did it come about?
The union was originally founded in 1992, but due to a combination of technical and political factors it failed to consolidate itself at the time as an industry union. There was a fundamental legal obstacle: historically, software was not recognised as an industry in Argentina, but rather as a service activity or administrative support. Only in recent years, after a lengthy legislative process, did the sector formally obtain industrial status — which is crucial because it radically changes our legal position and our capacity to represent workers before the State and companies.
On top of this technical difficulty, there were political reasons why we were not granted formal union recognition (personería gremial). To understand this, one must explain the Argentine union model, which differs from the European one. In Argentina you can have as many unions as you like (what is called inscripción, or registration), but the one that holds the exclusive right to bargain collectively and sign agreements is only one: the one with personería gremial, by virtue of being the most representative in its field.
For example, the Unión Obrera Metalúrgica (UOM) — the historic heavy industry union — sets the basic working conditions for all metalworkers in the country. The same applies in sectors such as food, construction, and transport. With a few specific exceptions, such as in some areas of the public sector where two unions hold personería gremial (ATE and UPCN), the general rule is that there is a single voice per sector. That was not the case in computing. There were two or three small organisations with a formal existence, but with no real bargaining power. We only received recognition of our majority representation and were granted personería gremial — which is essentially the legal right to represent workers before employers — very recently.
However, that recognition came in a very unstable political context. It happened during the final stretch of the center-right government of Mauricio Macri; it was a contradictory moment in which one part of his Ministry of Labour gave us its approval, while another sector of the same government, closer to the interests of the major technology companies, fought us through the courts. That left our legal situation trapped in an administrative limbo. Then, with the arrival of Alberto Fernández’s peronist government, there was an expectation of progress, but the process was slow and ultimately stalled in the state bureaucracy without being fully resolved. Now, under the current government of Javier Milei, we are directly confronted with a policy of demolition and destruction of the entire traditional union model.
At this moment we find ourselves in a situation where, legitimately and legally, we remain the recognised representatives, but we face far greater obstacles. Holding personería gremial is the first step, but the second is convening the collective bargaining table to sit down and negotiate a contract covering the whole country. To do that, you need a defined employers’ counterpart that is willing to sit down; but with four different employers’ chambers and a government that has halted all such dialogue processes, progress towards a national agreement is at a standstill.
Faced with this paralysis on the part of the State, our current strategy is to represent workers company by company. Instead of a grand national agreement that the government is blocking today, we sign specific agreements with individual firms. The law does not prohibit this: while the regulations state that the union with personería gremial represents the entire sector, they do not prevent us from reaching particular agreements with companies that choose to acknowledge their workers’ reality and sit down to negotiate.
And why is the national negotiation not moving forward? Is it a lack of political will from the government? How has the computing sector changed since this government took office?
Exactly. The government of Javier Milei, by political and ideological design, does not want national collective bargaining to exist. They can block it without facing major consequences: the Executive simply does not convene the parties and the courts do not compel it to do so. Legally, the situation is a disaster, but it is the reality we face today. It is a political game of attrition and deliberate destruction of union structures in favour of total deregulation.
Since this government took office, the situation has deteriorated noticeably. Because there is no National Collective Agreement establishing a floor of rights for everyone, the union can only exert influence in those companies where we already have a presence and specific agreements. In those places, conditions are better because we demand higher wages and negotiate every pay rise. However, in the rest of the industry, the reality is very poor: real wages are falling and paritarias — those annual wage negotiation rounds — have practically disappeared or are proving insufficient. Added to this is the fact that the Argentine economy is in a deep crisis and the sector is no longer growing at the rate it had been for the past five years. Up until 2024, growth was steady, but in 2025 the activity slowed down and began to plateau: one month there are layoffs, the next it picks up slightly, but the growth trend has been broken.
The international outlook has also become hostile. Even before the rise of AI, we already had the problem of rising interest rates in the United States, which triggered a contraction in risk investment, as we saw with the Silicon Valley Bank crisis. Argentina has the United States as its main export market thanks to the high calibre of our professionals, strong English-language proficiency, and time zone alignment, but while the domestic market is sinking in peso terms, external demand has been hit by these global financial shifts.
In this complex context, AI appears as an aggravating factor in the risk of layoffs. We are having very live debates about companies like Mercado Libre, which operates here as a monopoly: last year they laid off workers in areas that were replaced by AI-trained processes. There is a great deal of anxiety; there is some truth and some myth in the corporate narrative, but the climate of layoffs is real. Recently, companies such as PwC and Santander Tecnología have let go of hundreds of people from their technical divisions. Even in the banking sector there are reports of failures in money transfers following mass redundancies, leading people to think that attempts are being made to replace skilled staff with bots that do not yet work properly.
There is a very powerful campaign, led by figures such as Marcos Galperín (founder of Mercado Libre), arguing that there is no longer any need to study programming or computing because “AI does everything” or because now everything is vibe-coding. They suggest that you can hire any administrative worker with a chatbot subscription and be done with it.
From the union’s perspective, and in full agreement with academia, experts, and professional computing communities, we say that is absolutely false. On the contrary: AI creates new problems and more complex technical challenges that require more training, not less. Moreover, if companies are laying off junior and mid-level developers today, where will tomorrow’s senior engineers come from? Nobody will understand how the underlying systems work. Professionals need continuous upskilling, reading papers, and understanding new metrics. Without that formation, the industry has no future, and that is something we must work on together.
And who can join the union?
The union represents workers in the private sector whose primary activity is computing. In Argentina, classification depends on the activity of the company: if you are an IT professional but work for an insurance company, you belong to the insurance sector union; if you are an IT professional at a bank, you fall under the banking union; and if you are an IT professional working for the State, you belong to the public sector unions. But if you work for a company whose core business is software or hardware, then your place is our union. We are talking about companies such as Globant, Accenture, and Mercado Libre, as well as a great many consultancies and IT firms, both Argentine- and American-owned.
Besides wage bargaining, you mentioned that the union has a very strong social dimension. How does that work?
Exactly, collective bargaining does not stop at wages. The Collective Labour Agreement (Convenio Colectivo de Trabajo) is a general framework that regulates your entire working life. For example, in our agreement we include professional career development: we establish that if a worker specialises in a programming language or masters English, they must receive a bonus for that training. Rights such as special leave or holiday arrangements that improve on the general statutory provisions are also defined.
But beyond the workplace, unions in Argentina fulfil a social role that is quite distinctive. On one hand, we manage obras sociales, the health funds that are the backbone of the healthcare system for workers. On the other hand, we are key players in recreation and tourism. It is very common for unions to own recreational grounds, sports facilities, or even hotels. In Argentina it is perfectly normal to go on holiday to a hotel on the coast or in the mountains and find that it is owned by a union. Many unions also run their own football clubs or vocational training centres. For example, if a sector needs lathe operators, the union sets up its own training centres for workers. Sometimes the building belongs to the City of Buenos Aires and the teachers are public employees, but the union is the one that organises the training and supplies the students based on the needs of its industry.
All of this social welfare is usually enshrined in collective agreements. The worker makes a contribution from their salary to the union, and the union returns that in tangible services: free accommodation during holidays, access to campsites and swimming pools, or the provision of school bags and supplies for members’ children at the start of each school year. The union, in short, accompanies the worker across many more areas of their life, not only when there is a pay dispute.
As an IT union, we believe we have an active responsibility to engage in the debate about the platform economy. It is a topic that partly goes beyond our direct remit — for example, Uber workers are not strictly IT professionals — but we understand that our contribution from an academic and intellectual standpoint is very valuable. That is why we are driving forward the Observatory and the Laboratory: to tell the CGT, government officials, and members of parliament that we have a technical advisory capacity. Personally, for example, I have advised the Peronist labour commission on behalf of the union, and I was part of the advisory group for the artificial intelligence bill that was discussed at the time.
We do this technical work because we believe the union is a fundamental political actor in public debate. In Argentina this has a very strong tradition: in Perón’s era, more than 33% of public offices were held by union representatives. And in that spirit, we engage in the debate about digital sovereignty in the country. We defend the need for a state-owned enterprise to exist and for local data centres to be in place to safeguard personal data and the sensitive areas of the Argentine State, while also seeking greater operational efficiency.
We were directly involved in this. Some of our comrades were part of working groups in the previous government alongside Luis Papagni (Undersecretary of Administrative Innovation during Alberto Fernández’s government). The union played a role in what we believed was digital sovereignty: previously, Argentina’s digital management systems were in the hands of a private company, and when Luis came in, what was done was to bring those systems under the national State so that the State itself could manage them — making both the systems and the data centres part of the public domain.
As a union we are firmly convinced of the need for digital sovereignty in Argentina. Furthermore, historically, this union has championed free software and has driven an international free software conference that has hosted very prominent figures in the sector. In fact, we have an agreement with the Linux Foundation and have participated in international open source spaces such as OSPO. We are believers in that philosophy and integrate it into our technical and political vision.
And what do the Observatory and the Laboratory actually do?
The Observatory began, above all, out of a need to shed light on wages. In Argentina there is very powerful corporate propaganda claiming that software jobs are extraordinarily well-paid and that salaries are extremely high. I started looking for reliable sources because companies were citing surveys from employers’ chambers that were not public, and whose figures made no sense when we compared them with reality.
So we said: let us talk to our workers. And we began building our own data, using official sources — such as information from the state revenue collection agency — alongside other sources. From there, we developed a systematic body of work on wages in the sector.We then added another focus: in Argentina there is a promotional regime for software companies which, in practice, means they pay very little tax. So we began to show that these are leading companies, among the most valuable in the country, that nonetheless pay no taxes.
We began to highlight that many of these companies are registered abroad, in tax havens; that they keep their workers outside collective labour agreements; that they pay relatively low wages; and that, at the same time, they receive economic benefits from the Argentine State. From there we began producing more comprehensive reports on the state of the industry: approximately 160,000 workers, of whom 70% are men; age distribution; geographic location; working conditions; and above all the median real wage, which today sits at around one and a half million pesos ($1,500,000) — which is low.
By contrast, in the companies where we have a collective agreement, wages reach approximately 2,700,000 pesos, which is significantly higher. We also analyse how many workers are freelancers, how many are employed on a permanent basis, and how many are paid in dollars versus pesos. All of that work had an impact. It was picked up by national media, and companies such as Mercado Libre were forced to respond. There was an impact in government and in Congress: members of parliament began discussing the possibility of removing those tax benefits. In the end, they were not removed. During Alberto Fernández’s peronist government, there was scope to act, but no progress was made. And with the current government of Javier Milei, which is very closely aligned with the tech business sector, that possibility is practically non-existent. 
What we did achieve was placing the issue on the public agenda. We managed to make it a recurring subject of debate: so that whenever Galperín weighs in on social media, questions about the subsidies his company receives arise; and periodically the media returns to the topic. In other words, the policy was not changed, but a sustained challenge was generated.
The Laboratory, on the other hand, emerged from the need to intervene on a technical level. We have a project for real-time verification of platform algorithms, with the aim of detecting possible violations of labour regulations: unrecognised overtime, payment problems, wage discrimination, or forms of surveillance extending beyond the workplace. The idea is to develop a platform or verification system. That work has already begun. We also have a preliminary version — a proof of concept — that we developed together with the Universidad de San Martín, funded as a research and development project, called VerifyLabor.
One of the objectives is to advance that system. Other objectives include organising hackathons, producing publications, and curating technical articles, especially on artificial intelligence and its implementation. We also aim for the Laboratory to be able to develop its own positions on technical debates — for example, the question of technical debt, which is also a political discussion. Looking ahead, we are thinking about organising conferences and other activities that may emerge from that space. That, broadly speaking, is the objective of the Laboratory. And together with the Observatory, because we conceive them as parts of the same project: the Observatory with a more research-oriented focus, and the Laboratory with a more technical and development-oriented one.
And what is the ultimate goal? Would it be to build an international network?
Yes, but it is an objective of the union as a whole, not just of the Laboratory or the Observatory. The idea is to build an international union capable of engaging in these debates in spaces such as the ILO or at global conferences, and with the capacity to address companies like Google, Anthropic, and others directly. To be able to say that there are limits, that not everything is permissible — especially when it comes to working conditions.
It also involves discussing more structural questions: if technology is going to make workers more productive, then why does that not translate into better wages or a reduction in working hours? And, moreover, putting the ethical questions linked to artificial intelligence on the table. Today that discussion is completely dominated by the companies themselves. It is presented as a debate between actors such as OpenAI and Anthropic, but in reality the “ethics” they put forward tends to be aligned with geopolitical interests: basically, as long as the systems work against the enemies of the United States, anything goes. In that scenario, a real counterpart is missing. What is missing is for IT workers to be represented in that discussion and to be able to intervene from their own perspective.
I was thinking about the case of Mercado Pago in Argentina and its comparison with the Pix system in Brazil: how do you see that relationship and what differences seem most relevant to you?
In Argentina there is a comparable local example, which is Cuenta DNI. It works with your national identity document: with your DNI you can already make transfers. It is an application from Banco Provincia. It offers discounts and cashback, works extensively with neighbourhood businesses, and has enabled fairly widespread digital inclusion. In practice, you have an account and a payment app, similar to Mercado Pago, but linked directly to your identity.
It is not a Central Bank system like Pix in Brazil — it belongs to a provincial bank. But it is interesting for a legal reason: Banco Provincia predates the National Constitution, so it has a presence across the entire territory and, because it is a digital wallet, it can operate throughout the country. And now there are discussions about Pix beginning to be used in other countries as well. That is why we argue that there should be a nationally owned tool that competes with both Cuenta DNI and private digital wallets. There is also MODO, a wallet promoted by several private banks, but it does not have the same benefits scheme or the same impact in terms of financial inclusion.
And the pan-American perspective? Do you speak with unions in other countries?
Dialogue with other countries does exist; there are regional exchange forums. But the reality is that there is no Latin American bloc that functions in a coordinated way. In Argentina there is Mercosur, even the Mercosur Parliament, but in practice it has very little influence.
There was a period when greater regional integration seemed possible — during the governments of Lula in Brazil, Chávez in Venezuela, and Morales in Bolivia — but that process did not consolidate. Today the situation in Latin America is quite fragmented: each country moves forward on its own, with a few specific exceptions. And this is not purely an ideological matter. There was no significant integration even between governments of a similar political sign, such as Macri’s in Argentina and Bolsonaro’s in Brazil. It is one of those processes that everyone considers important but that has never managed to develop in practice. The most that exists today is Mercosur, which survives, but functions essentially as a trade agreement, without any real political or technological integration.