I don't know if I'm free in a world like that
Numa manhã de debate em Lisboa, uma sala de profissionais entrou convicta de que a inteligência artificial não representava ameaça real — e saiu com uma visão diferente. Na Conferência de Sustentabilidade do Negócios, na NOVA SBE, dois interlocutores colocaram em confronto não apenas tecnologia, mas filosofias opostas sobre poder, liberdade e responsabilidade democrática. O que ficou suspenso no ar não foi uma resposta, mas uma pergunta cada vez mais urgente: a quem respondem os sistemas que passamos a obedecer?
- Uma audiência inicialmente cética foi confrontada com a ideia de que delegar decisões críticas a sistemas privados pode esvaziar silenciosamente a soberania dos Estados e a autonomia dos cidadãos.
- A tensão central do debate não era técnica mas política: quem constrói estes sistemas, quem os governa e quem responde quando falham?
- Stephan de Moraes defendeu que os ganhos de eficiência e a democratização do conhecimento superam os riscos, e que os contrapesos regulatórios existentes impedem qualquer concentração absoluta de poder.
- Adolfo Mesquita Nunes, assumindo-se liberal clássico favorável à iniciativa privada, disse encontrar-se a defender regulação — e alertou que esse paradoxo deveria dizer algo ao público.
- No final, a maioria da sala votou que a IA representa uma ameaça real, sinalizando que os argumentos sobre soberania e prestação de contas democrática ressoaram mais fundo do que os presentes esperavam.
A sala começou a manhã sem grande alarme. A inteligência artificial era vista como uma ferramenta promissora, não como uma ameaça. Mas o debate que se seguiu na Conferência de Sustentabilidade do Negócios, na NOVA SBE, acabaria por inverter esse consenso.
Adolfo Mesquita Nunes, advogado na Pérez-Llorca, não se apresentou como tecnófobo — pelo contrário, reconheceu o potencial transformador da IA. O que o preocupava era outra coisa: a transferência estrutural de poder para sistemas controlados por empresas privadas sem qualquer mandato democrático. Quando domínios como a defesa, a justiça ou a administração pública passam a depender de entidades que não respondem perante nenhum eleitorado, algo fundamental muda. "Não sei se sou livre num mundo assim", disse, apontando para um futuro em que algoritmos medeiam o acesso à informação e moldam decisões que afetam a vida quotidiana.
Stephan de Moraes, da Indico Capital Partners, colocou a IA ao lado da descoberta da eletricidade em termos de impacto civilizacional. Falou de eficiência, de libertação do trabalho repetitivo, de uma democratização sem precedentes do conhecimento. Quanto à concentração de poder, argumentou que os mecanismos regulatórios existentes são suficientes para evitar domínios absolutos — e que o poder está, afinal, mais distribuído do que a narrativa do medo sugere.
O debate nunca chegou a acordo, mas produziu algo mais revelador: moveu a audiência. Os mesmos profissionais que tinham chegado dispostos a desvalorizar o risco saíram convictos de que ele existe. A jornalista Helena Garrido moderou uma troca que, no fundo, não era sobre tecnologia — era sobre quem decide, quem responde e a quem servem, afinal, as inteligências que estamos a construir.
The room began the morning convinced that artificial intelligence posed no real threat. By the time the debate ended, a majority had changed their minds entirely.
The session, held at the Negócios Sustainability Conference at NOVA SBE, was structured as a direct confrontation between two visions of AI's future. On one side sat Adolfo Mesquita Nunes, a partner at the law firm Pérez-Llorca, tasked with making the case that AI represents a genuine danger. On the other, Stephan de Moraes, managing general partner of Indico Capital Partners, argued that the technology's benefits far outweigh its risks. Journalist Helena Garrido moderated the exchange, which proved more persuasive than either speaker may have anticipated.
Nunes began by rejecting the label of technophobe. He called himself an enthusiast for technology's possibilities—the millions of opportunities AI could unlock. But he drew a sharp distinction between opportunity and threat. The real danger, he argued, lies not in what AI can do but in the structural transformation it forces upon society. When governments, corporations, and citizens delegate their power to systems controlled by private companies answerable to no electorate, something fundamental shifts. The threat grows proportionally to the power surrendered. Critical domains—defense, justice, the administration of law itself—could come to depend on entities that have no obligation to the public. "I don't know if I'm free in a world like that," Nunes said, pointing to a future where algorithms mediate access to information and shape the decisions that affect daily life.
Moraes acknowledged that all major technological transformations carry some risk. But he positioned AI alongside the discovery of electricity in its transformative scope—perhaps even more profound. The technology increases human efficiency, frees people from repetitive labor, and democratizes access to knowledge in ways previously unimaginable. He spoke of an enormous equalization happening through the distribution of intelligence itself. Regarding the concentration of power, Moraes argued that sufficient counterweights exist. Governments and regulators have tools to prevent absolute dominance by any single actor. "Power is far more distributed than people think," he said, rejecting the notion that tech giants are destined to become the world's new masters. He pointed to concrete opportunities: new materials, energy solutions, personalized medicine, responses to aging populations.
The disagreement centered ultimately on governance and accountability. Nunes, describing himself as a classical liberal who favors private enterprise, said he found himself defending regulation—a position he noted should signal something important. The goal was not to halt AI but to ensure it served people rather than the reverse. Who builds these systems? Who decides how they operate? Who answers when something goes wrong? These were not technical questions but political ones.
Moraes countered that existing institutions could manage these concerns. The debate never resolved into agreement, but it did something more consequential: it moved the room. The audience that had begun the morning skeptical of AI as a threat voted at the end to affirm that it was one. The shift suggested that Nunes's warnings about sovereignty, autonomy, and democratic accountability had found purchase with listeners who came prepared to dismiss them.
Notable Quotes
The threat grows proportionally to the power we delegate to artificial intelligence in terms of sovereignty, autonomy, and knowledge— Adolfo Mesquita Nunes
We're talking about an enormous equalization through the democratization of intelligence— Stephan de Moraes
The Hearth Conversation Another angle on the story
Why did the audience change its mind so dramatically? What did Nunes say that Moraes didn't?
Nunes made it personal. He didn't argue that AI would fail or that the technology was flawed. He asked: who controls the system that controls your life? That's a question about power, not capability.
But Moraes had an answer to that—he said regulators and governments would keep things balanced.
He did. But Nunes was asking something deeper: can you actually regulate a system you don't understand? And if the companies building it don't answer to voters, does regulation matter when it's too late?
So the threat isn't AI itself. It's who owns it.
Exactly. Nunes said he loves technology. He just wants to know who's in the room when decisions get made. Moraes said that's already distributed. But the audience seemed to think: maybe not distributed enough.
Did Moraes lose because his argument was weak?
No. His argument was strong—efficiency, democratization, medical breakthroughs. But he was answering a different question. He was saying "this will be good for you." Nunes was asking "who decides what good means?"
And that question scared people more than the promise of benefit.
It did. Because once you ask it, you can't un-ask it. You start noticing how many decisions you've already delegated without knowing it.