An outburst that went viral
A firsthand account published on social media by a British researcher has reignited this week one of the thorniest debates in contemporary technology: how far is it worth going in building ever more powerful artificial intelligence systems?
Last Tuesday (8/9), Jacob Coxon drew the internet's attention when he publicly announced his resignation from two of the sector's most influential companies — OpenAI and Anthropic — after years dedicated to developing AI models. The announcement quickly spread across digital platforms and reignited discussions about ethics, transparency, and responsibility in the industry.
Who the researcher is
Coxon is British and built his career right in the heart of the industry he now criticizes. His trajectory includes stints at leading laboratories competing for global leadership in creating generative systems — the same circuit responsible for products that, in recent years, have transformed the way people write, program, and research.
It was within this environment that the researcher says he reached an uncomfortable conclusion: the pace of innovation is far ahead of society's ability to understand and regulate it.
The warning about "superintelligence"
The central point of his statement is the concept of superintelligence — a hypothetical form of artificial intelligence capable of surpassing human performance in virtually every field of knowledge.
According to the researcher, the problem lies not only in distant science-fiction scenarios, but in the very speed at which teams are pressured to release more capable models. For him, this accelerated advance represents a danger without parallel in history, because decisions made today can have irreversible consequences.
- The concern is not limited to technical errors, but to the absence of solid control mechanisms;
- The researcher argues that laboratories should review their priorities and exercise greater caution before releasing more powerful systems;
- He also calls for more transparency about the risks identified internally by companies.
The weight of speaking from the inside
Accounts like Coxon's gain strength precisely because they come from someone who knows the backstage. In recent years, professionals linked to major AI laboratories had already left their posts citing ethical dilemmas, which fuels a pattern of distrust: if the creators themselves are warning, what does that say about what is being built?
For experts in technology governance, the episode highlights a structural tension in the sector — commercial competition between companies pushes development forward, while safeguards tend to move at a slower pace.
The debate that should not end
Coxon's statement does not offer definitive solutions, but it works as a call for public debate. Between defenders of rapid advancement, who see gains in health, science, and productivity in AI, and critics who demand brakes and independent audits, the possible consensus may lie at an intermediate point: openly discussing the risks before they become irreversible.
What the case makes clear is that the question about the limits of artificial intelligence is no longer merely technical. It is political, ethical, and increasingly urgent.