Leaders in the artificial intelligence industry, including executives from companies like Anthropic, OpenAI, and Google DeepMind, have been warning about the risks of advanced AI systems escaping human control. Dario Amodei, CEO of Anthropic, the company responsible for developing Claude, stated that a combination of increasingly powerful AI agents could, in a period of just six months to one year, take control of significant portions of the internet.
The global technology race, driven by the competition between the United States and China for leadership in AI development, has increased pressure for rapid advances, even in the face of security concerns. Researchers and executives argue that the speed of development is outpacing the ability of governments and companies to understand and control its potential effects, calling for independent safety evaluation mechanisms.
There is no consensus among specialists on the possibility of AI representing an existential threat to humanity. Former Anthropic security researcher Jacob Coxon estimated a 10% chance of artificial intelligence causing human extinction within a decade, although this assessment is far from representing a scientific consensus.
Given this scenario, pressure is growing for governments to establish international rules that keep pace with technological evolution. The challenge is to find a balance between innovation and security, since overly rigid regulation could reduce competitiveness, while the absence of rules could allow powerful systems to operate without sufficient control mechanisms.
