'Traditional unions struggle to understand tech sector'

In this episode of our “Political Construction of Artificial Intelligence” series, we feature an interview between Diyar Saraçoğlu and Simone Robutti from the Tech Workers Coalition (TWC), an organization engaged in labor organizing across the US and Europe.

Robutti, a programmer and machine learning engineer, conducts research on the social harms of artificial intelligence. As one of the founders of TWC Berlin and TWC Italy, he has played an active role in worker organizing efforts and is currently involved in TWC’s work on infrastructure, education, and the development of new chapters.

The interview explores how tech workers are organizing, the impact of AI and automation on labor movements, and the dynamics shaping international solidarity. It also highlights TWC’s broader mission—not just improving working conditions but transforming the tech sector into a more just and democratic space. Robutti’s insights provide a valuable perspective on the current state of tech labor and its evolving landscape.

'TWC trains workers in fundamental organizing methods'

How do you evaluate the organizational dynamics of workers in the technology field? Can we speak of an evolving form of organization over the years?

This varies by region—organizing in the USA differs from Italy, for instance. But broadly speaking, the Tech Workers Coalition (TWC) focuses on established organizing practices rather than reinventing the wheel just because tech workers feel “special.”

TWC trains workers in fundamental organizing methods and emphasizes that traditional approaches still work. However, remote work has introduced new challenges. Many established organizing strategies assume workers meet in offices or communities, which doesn’t apply to many tech companies. There’s no universal solution, but we try to meet workers in digital spaces outside company surveillance—just as in the real world, where you’d avoid meeting under a manager’s watch.

This can be trickier than inviting colleagues for a drink; asking someone to chat on a different platform instead of Teams feels unusual, but it happens and works. Another approach, especially in large companies, is engaging workers in their existing online communities, such as company-specific subreddits or Facebook groups. While TWC doesn’t do this directly, others have found it useful, particularly for global coordination, as seen in movements like No Tech for Apartheid.

For smaller companies with physical offices, traditional unions still play a central role, with TWC acting more as a complementary force. Some new unions, like UTAW in the UK—co-founded by TWC members—have emerged, but they don’t necessarily promote distinct organizing models.

One key difference in tech organizing is strike strategy. Short strikes have little economic impact since lost work can be made up later. They function more as warnings unless they damage the company's public image. Longer strikes can be more disruptive; for example, after a one-week strike at The New York Times, system issues began to emerge. However, it remains uncertain how long it takes for digital infrastructure failures to cause significant economic harm.

Another experimental strike method, seen in Italy, involved consultants refusing work for two hours at the end of the day and setting an auto-reply stating they were on strike. This maximized pressure on the company by making the disruption visible to clients and also prevented unpaid overtime. The tactic was highly effective, forcing the company to concede in just two days.

Ultimately, tech organizing is still evolving, with workers adapting traditional methods to the unique challenges of the industry.

'Company size matters in unionization'

What unique qualities do unionization and organizational practices in the technology sector possess compared to other sectors? What similarities and differences emerge in this regard?

Geographical differences play a major role, often aligning with company structures—corporate, startup, or consultancy—since work dynamics vary significantly.

For example, in Italy, the workforce is highly fragmented due to numerous small and medium-sized consultancy firms working for the same client. Each company might assign only a few workers to a project, scattering employees across different locations. Since workers serving the same client have different legal employers, it's unclear whom to address with demands, making organizing particularly challenging. This dynamic doesn’t exist in the same way in the US.

In the US, most organizing efforts focus on small to mid-sized startups (50–500 employees) or on giants like Google, Amazon, and Facebook. Ultimately, company size often matters more than the sector itself, as organizing practices differ vastly between small startups and tech giants. There’s no clear theoretical framework for sub-sectors in tech production, and unions generally treat the industry as one broad category.

Localized strategies

As the Tech Workers' Coalition (TWC), could you elaborate on the organizational and solidarity activities you conduct? What strategies and methods do you employ to implement your initiatives?

Given the fragmentation of the tech workforce, each TWC chapter develops its own strategy based on local conditions and legal frameworks. For example, organizing in Italy differs greatly from the USA.

That said, a common theme across chapters is TWC’s complementary role to unions, with a primary focus on education and communication. We connect workers with unions, ensuring they are already informed, engaged, or trained in organizing. This takes various forms. TWC Global, for instance, runs structured training programs like “Powering On”, while in Berlin, efforts center on Works Councils—a distinct institution in Germany, Austria, and the Netherlands.

Works Councils provide democratic worker representation within companies but are limited in their ability to engage in direct conflict, such as striking. However, they have legal rights, including access to company finances and the ability to prevent unjust firings. Since forming a Works Council requires less effort than unionizing, the main challenge is ensuring pro-worker candidates win elections rather than company-aligned ones.

To support this, TWC Berlin provides training, assists with campaign materials, and offers legal guidance on complex regulations. We maintain a Telegram chat for knowledge-sharing and organize an annual conference for Works Council members from various tech startups. The last event gathered 200 elected representatives, a significant achievement in this context.

These strategies are highly localized—what works in Germany wouldn’t necessarily apply elsewhere, underscoring the need for flexible, region-specific approaches.

'Lack of unionization examples is a major barrier'

What obstacles do you encounter when bringing together technology workers dispersed across different countries to create a shared sphere of action? How do you overcome these challenges?

Remote work is one challenge, but a bigger obstacle used to be the stereotype of tech workers as anti-union individualists. This perception has shifted. The workforce is now more diverse, with more women and people of color, particularly in the US, and mass layoffs have changed the overall mood. While tech workers aren’t as union-ready as dockworkers, they are far more open to organizing than in the past.

A major barrier today is the lack of unionization examples. Many workers are theoretically interested but lack both practical experience and role models. Yet, when workers do take initiative, they often succeed. The biggest challenge is taking that first step. When union efforts fail, it’s usually due to aggressive union-busting, which has intensified. Years ago, tech companies were caught off guard by unionization; now, they are well-prepared and trained to counter it.

Another issue is that traditional unions often struggle to understand the tech sector. This is especially true in Italy, though the situation is slightly better in the US and Germany. Many unions fail to grasp the complexities of tech roles, which can alienate workers. For example, not knowing the difference between a developer and a tester might seem trivial, but tech workers value specialization.

This gap creates practical challenges. I once spoke with a union organizer from Italy’s CGIL, experienced in automotive factories, who asked if moving a worker from development to testing constituted a demotion (demansionamento). In a factory, demotion is clear-cut; in tech, it's more ambiguous. These nuances highlight why many unions struggle to adapt, and why we focus on bridging this gap at a local level.

'Tech executives are increasingly fearful'

At Donald Trump’s inauguration in the US, technology giants were seen side by side. In your opinion, does this signal the onset of a new era in the technology sector? What do you think awaits technology workers under these new conditions?

Yes, absolutely. We were prepared for this. When Trump was first elected, many tech workers joined TWC, and this time even more are mobilizing. The illusion that the tech sector is a force for good has collapsed. The old political alignment with American liberals is gone, leaving many workers disillusioned. Simple dissent is no longer enough—conflict is inevitable, and both sides recognize this.

Tech executives are increasingly fearful. In a recent interview, Marc Andreessen admitted they were “days away from riots” and described tech workers as going “feral”—a paranoid exaggeration, but revealing of the growing tension. This fear signals more resistance from management, but also makes it easier to convince workers to unionize. However, organizing will likely become even more difficult, especially in the U.S.

In Europe, distrust toward American tech is growing, particularly after recent Starlink controversies. While skepticism about U.S. infrastructure isn’t new, it's now entering mainstream political discourse. This could push for more autonomous tech development in Europe—but whether that materializes remains to be seen.

Effect of Palestine conflict

Technology workers not only come together to defend their own rights but also engage in political activism—for instance, showing solidarity with Palestine. How do you assess the impact of this diversity on the potential for organization within the technology sector?

This is a crucial topic, and as TWC, we fully welcome the politicization of worker power. Our goal isn’t just to improve working conditions but also to channel that power into broader political struggles. While groups like No Tech for Apartheid focus on single issues, their work still strengthens collective action.

Political and social issues often drive tech workers to organize. During Trump’s presidency, many joined TWC in response to workplace discrimination and harassment—not from a traditional leftist perspective, but because they saw firsthand how exploitation affected them. The Palestine conflict has reinforced this realization: workers see that their managers and companies are complicit in harmful systems, fueling distrust and mobilization.

These initiatives have tangible effects. While major companies rarely cancel contracts, smaller firms do, and the broader impact is the growing awareness that tech workers have a say in how their labor is used. Many are now questioning their industry's role in oppression and realizing their collective power. The hard wins are contract cancellations; the softer but equally important wins are the increasing number of workers seeking ways to resist.

Tension between union organizing and cooperatives

Do technology workers have the potential to harness the power of production processes to develop self-organizing practices that challenge the fundamental dynamics of capitalism? Are there any projects you are involved in regarding autonomous production cooperatives, worker-controlled software projects, or other collective organizational models?

TWC doesn’t actively focus on autonomous production cooperatives or worker-controlled software projects, though we see them as part of a broader vision. If we succeed in building worker power within companies, the long-term goal is to encourage alternative ways of producing technology—but we’re not there yet.

Currently, tech cooperatives and other collective models aren’t a strategic priority for TWC, though we engage with spaces that support them, such as NGOs, open-source projects, and study groups. Some members, including myself, have experience in cooperatives, but this remains peripheral to our main work.

There’s also a tension between union organizing and cooperative efforts. Historically, those who leave companies to form cooperatives don’t contribute to workplace struggles, whereas those who stay help build unions and push for systemic change from within. This pattern holds true in tech as well.

That said, alternative organizational models, particularly those inspired by agile methodologies and sociocracy, are influencing cooperative structures. There’s a growing exchange between the tech sector and democratic organizational practices, shaping how new worker-led initiatives might emerge in the future.

A critical stance on generative AI

How are emerging large language models (LLMs) and artificial intelligence technologies shaping the organizational processes of technology workers?

As far as I know, LLMs and AI technologies have not significantly influenced tech worker organizing. TWC takes a largely critical stance on generative AI, and I share that view. Even though I’ve organized conferences on potential leftist uses of AI, nothing particularly promising has emerged.

I haven’t seen any meaningful applications of AI in organizing, even at an experimental level. Some attempts exist, like politicized chatbots—one friend created an anarcho-communist chatbot designed as an ant from an ant nest, using retro video game aesthetics to spread anti-work and union rhetoric. However, this remains more of an artistic experiment than a practical tool.

While some union organizers might use LLMs for drafting manifestos, communications, or emails, there’s no structured or transformative adoption of AI in organizing efforts yet.

LLMs like DeepSeek, developed in China, have recently gained popularity across various social media platforms and shopping sites. Clearly, despite various obstacles, technologies developed by China-based companies are set to become even more popular. What do we know about the working conditions of technology workers who develop these tools in China (and in countries outside of the US and Europe)? Do you have any established contacts in this area?

TWC previously had a chapter in India, which eventually transitioned into a union. However, we have no direct connections with tech workers in China, though we would love to. Reaching out is difficult due to language barriers, censorship, and the underground nature of labor organizing there. There’s little visibility into their conditions, and no established bridges exist, as far as we know.

In Latin America, we haven’t seen major developments. Despite a strong Brazilian presence in our Berlin networks, no significant tech labor movement has emerged there yet. Info Proletarios exists, but it’s not substantial at this stage.

While we’re eager to connect with Chinese tech workers, there’s currently no infrastructure or platform for meaningful exchange.

'Urgency to organize is greater than ever'

In an era where artificial intelligence and automation are becoming increasingly widespread, what strategic approaches do you anticipate will be most effective in advancing the rights and demands of technology workers in the future? Do you think technological advancements will result in fundamental changes in organizational forms?

The current AI hype won’t last forever, and the narrative of AI replacing highly skilled workers will likely fade as well. AI isn’t firing workers—managers are. AI simply serves as an excuse, and even if that excuse disappears, working conditions will still deteriorate.

Most digital automation, particularly AI and machine learning, doesn’t eliminate work but shifts it to a more precarious, globally fragmented workforce. Instead of replacing programmers, AI trends push more tasks onto low-paid data labelers. This pattern isn’t new—offshoring tech work to lower-wage countries has been happening since the 1990s with advancements like Java, which made geo-arbitration easier. AI is just the latest justification.

The key strategy is to unionize as quickly as possible. Tech workers are losing their privileged status, and this decline must be leveraged while it still holds weight. Meanwhile, the new wave of tech workers—often from bootcamps or with fewer protections—will be easier to organize due to their material conditions.

The biggest shift isn’t automation itself but capital’s growing confidence in cutting jobs outright. Previously, senior workers were replaced with younger, cheaper employees. Now, companies are simply eliminating positions. This trajectory was always expected, but the sudden acceleration has been a wake-up call.

In places like Italy, where tech workers have always been less privileged than their American counterparts, the fear of losing what little security remains has been a strong motivator for organizing. In contrast, American workers often assumed their status was permanent—until mass layoffs shattered that illusion. Now, as job cuts intensify, the urgency to organize is greater than ever. (DS/VC/VK)