Anthropic: 'Uncontrolled AI could dominate the Internet in less than a year'

Dario Amodei, CEO of the artificial intelligence company Anthropic, has issued an unprecedented warning about AI development. In an interview with CBS News, Amodei stated that without the implementation of adequate safety measures, artificial intelligence could gain the ability to control the entire internet within the next 6-12 months.

Amodei described the technological development as an "exponential" progression, comparing it to a sequence that continuously doubles. He admitted he did not initially fully grasp the implications of such rapid progress, calling the current situation a "wake-up call" for the entire industry. The primary concern revolves around the possibility that AI could command swarms of autonomous agents capable of taking control of global digital infrastructure.

Amodei's warning comes in the context of unprecedented growth in artificial intelligence capabilities, a technological field that has seen dramatic acceleration in recent years. The starting point of this race can be traced to the first large-scale language models, which demonstrated increasingly sophisticated abilities to interact with digital systems. Anthropic itself was founded in 2021 by former OpenAI researchers, with the stated goal of developing AI "aligned" with human values and safe.

The path that led to this alarm is characterized by intense competition among technology companies seeking to develop increasingly powerful models, often without adequate safety controls. The progression described by Amodei—"one, then two, then four, then eight"—reflects how incremental improvements have created a cumulative effect that now threatens to escape human control. The possible consequences of this uncontrolled development include not only the risk of AI agents operating autonomously on the Internet, but also implications for global cybersecurity and the digital geopolitical balance.