2026-09-13 · Fuente:Tom's Hardware

Anthropic's CEO says an AI-driven botnet swarm could take over the entire internet within 6–12 months, while an ex-Anthropic figure warns AI will become self-sufficient and 'kill our posterity.'
Anthropic chief executive Dario Amodei has issued a stark warning about AI-driven cyberattacks, saying that swarms of AI-coordinated botnets could soon be powerful enough to take over the entire internet. "In 6–12 months such a swarm could be capable of taking over the entire internet with a persistent botnet," he cautioned, per Tom's Hardware.
The scenario he describes centers on a botnet "swarm" — networks of compromised machines orchestrated by artificial intelligence rather than human operators. In that model, AI would handle coordination, targeting and persistence autonomously, letting a malicious swarm entrench itself across connected systems at a scale and speed beyond conventional, manually run botnets. The stated timeline of roughly six months to a year frames the danger as near-term rather than theoretical.
The report also carries a parallel warning from an ex-Anthropic voice, who cautions that AI will become self-sufficient and "kill our posterity" — the risk that autonomous systems could ultimately extinguish humanity's descendants. Taken together, the two statements reflect concern emanating from inside and around the AI lab about systems that operate, adapt and perpetuate themselves beyond human oversight.
For the companies building and buying AI compute, the remarks land on territory they directly control: internet-connected infrastructure. A persistent, AI-run botnet capable of "taking over the entire internet," as the warning puts it, would implicate the data centers, networks and accelerator fleets underpinning the AI buildout. Whatever becomes of the 6–12 month timeline, the statement injects a security consideration into infrastructure planning that has largely revolved around supply, power and cost. For an industry whose spending decisions hinge on trust in connected compute, the warning is a reminder that capability gains cut both ways.
Etiquetas: Anthropic · Dario Amodei · AI Botnet · Cybersecurity · AI Safety
萬安算交所 AIXX · AIXX · Exchange verificado · Ir al mercado