Regulate or innovate? The debate at the STEM Women Congress Madrid 2026
The Ethics Cachroom debate at the STEM Woman Congress Madrid 2026 explored one of the defining tensions of today’s technological landscape: how to balance fast-moving AI innovation with the need for ethical safeguards.
On one side, those who argue that without regulation, artificial intelligence can become an uncontrollable risk. On the other, those who believe that regulating too soon can hold back progress and leave Europe behind.
The Team Pro Regulation set limits to protect people, arguing for mandatory frameworks to protect human rights, ensure transparency, and build trust. The team included Sorne Resines (Directora de Seguridad para España y México, Alpatech Core Services), Kelly Cuesta (Senior Marketing Consultant, Director and Founder, Latinas In Tech Spain), Olga Polo (Chief Compliance Officer, Abertis), Diana Martín (HR Manager Iberia), and Javier Espinosa García (Director of Sowtwer Engineering at CaixaBank Tech).
While the Team Pro Innovation give technology the space to advance, warning that excessive regulation risks slowing competitiveness, increasing costs, and pushing talent away from Europe. The team included Berta Tortajada (Senior Legal Counsel, Oracle), Juan José Cerrolaza (Head of Artificial Intelligence and Application Data and AI, Kyndryl), Marta Portalés (Head of European Project, Mobile World Capital Barcelona), Tanya Suárez (Founder and CEO, IoT Tribe), and Arantxa Aramendía (IT Project Manager, Volkswagen Group España Distribución).
The case for regulation: “protect people”
Team Pro Regulation stated that artificial intelligence needs rules, not as an obstacle, but as a foundation. Their case rested on three key pillars: avoiding automated decisions without human oversight, guaranteeing traceability, audit trails, and compliance, and building secure systems from the ground up. Their central message was that regulation is necessary because without regulation, AI can discriminate, leak sensitive data, and make decisions with real-world impact without clear accountability. They used concrete examples to show the risks: an employee asking an AI assistant about changing jobs could have that information exposed to HR; medical data could be used against people in insurance or hiring decisions before they even know it happened.
Without trust, there is no adoption and without adoption, there is no sustainable innovation. Anticipating regulation does not penalise companies, it positions them better against reputational, legal, and financial risk.
In the US, ZIP codes are used to determine bail conditions, systematically punishing communities like Harlem over Park Avenue. Amazon’s AI recruiting tool, now a well-known historical case, discriminated in favor of white men aged 35. Innovation without regulation scales fast and destroys trust. Without trust, there is no sustainable business. Regulation, they argued, is not a brake. It is a strategy. Anticipating it makes innovation sustainable. And predictability is what investors actually want.
The EU AI Act was presented as a way to set ethical limits for all companies operating in Europe. Their final argument was the “Brussels Effect”: when Europe regulates, global standards often follow, as happened with GDPR.
The case for innovation: Europe is falling behind
Team Pro Innovation argued that AI is not developing in a legal vacuum. Existing laws already apply, and legal systems have always adapted to new technologies over time.
On one hand, Europe leads on rights protection and sets global standards, as happened with GDPR. On the other, it lags in technological development and attracts less AI investment. In 2024, 35 billion dollars were invested in AI startups globally, only 6% of global investment in AI startups went to Europe.
Of the leading AI models in the world, 40 are American, 15 are Chinese, and 3 are French. A third of European companies report significant delays in product development due to regulation. Companies spend between 90,000 and 500,000 euros per year on regulatory compliance alone, and up to 60% have had to deactivate or withdraw product capabilities to meet requirements. The EU itself acknowledges this, estimating that overregulation costs each company over 500,000 euros per year.
They cited the case of Aleph Alpha, a German large language model company valued at 20 billion euros, acquired by Canadian company Cohere because it could not develop its models in Europe under current regulatory conditions. The GDP comparison landed hard: over the last ten years, income per person in the US grew by 15,000 euros, in Europe, it grew by 4,000.
The jury pushes back
Alberto Barroso brought data to the room: a Stanford report published two weeks before the congress showed a 55% increase in AI incidents in the last year. The four major providers, OpenAI, Anthropic, Google, and Meta, have stopped reporting critical details on training data, model weights, and parameters. Hallucination rates across the top 26 AI models range between 25% and 95%. Should regulation apply equally to large tech companies and small startups?
Natalia Álvarez raised the timing problem: AI advances faster than legislatures can move. How do you avoid regulating too late, when rules are useless, or too early, when they are harmful?
Ana Pérez brought the conversation back to people. A major e-commerce company had to pull its AI-based hiring system after it produced gender-biased results. If AI screening produces biased outcomes in the absence of regulation, what happens to female representation in STEM? And if regulation steps in, who is actually responsible for making sure it corrects the inequality rather than just documenting it?
What both sides agreed on
Both teams placed people at the center: Team Pro Regulation said anticipating regulation makes innovation sustainable, Team Pro Innovation said technology improves quality of life and must operate within a valid legal framework.
The difference, as Joana Barbani noted in her closing, was about speed, one side wants to move fast, the other one wants to move carefully.
Both sides ultimately support the same objective: human-centric artificial intelligence. The teams at STEM Women Congress Madrid 2026 showed that the most important conversations in technology are happening between people willing to argue in good faith, challenge each other with real data, and leave with more questions than they arrived with.