This new AI-powered computer worm can learn to attack any device - ThePrint

2026-06-08 · Source: artifical intelligence via Google News · Field: Technology & Digital — Cybersecurity & Data Privacy, Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

Researchers at the University of Toronto have developed an AI-powered computer worm capable of adapting to attack any online device, detailed in findings published on arXiv.org on June 2, 2026. Unlike traditional worms that exploit specific vulnerabilities, this new malware utilizes cheap and free-to-use AI models to learn and adapt its strategy, employing a recursive reasoning loop to detect and exploit diverse security flaws across networks. In a controlled lab simulation, the worm penetrated 75% of a 33-machine corporate network within one week, operating without human intervention. Once inside, it leverages the infected system's processing power for reasoning and target identification, drastically reducing the cost of cyberattacks to "nearly zero." This development, alongside concerns about models like Anthropic's Mythos, highlights a significant cybersecurity threat.

Key takeaway

For cybersecurity experts and policymakers evaluating network defenses, this adaptive AI worm signifies that traditional vulnerability patching is insufficient. You must prioritize developing dynamic, AI-driven defense mechanisms that can counter evolving, learning threats. Proactively engage with research on AI malware to understand its propagation methods and implement robust, multi-layered security architectures, as no system is inherently immune to these new, low-cost attack vectors.

Key insights

AI-powered worms can adaptively exploit diverse vulnerabilities, posing an unprecedented cybersecurity threat.

Principles

• Adaptive worms use recursive reasoning.
• AI reduces cyberattack costs to near zero.
• No computer system is immune.

Method

The AI worm operates by entering a network, then using the system's processing power for reasoning to identify and exploit diverse vulnerabilities in connected devices, propagating autonomously.

In practice

• Simulate AI worm attacks in secure labs.
• Consult national security bodies on AI threats.
• Patching specific vulnerabilities is insufficient.

Topics

• AI-powered Worms
• Adaptive Malware
• Cybersecurity Threats
• Network Security
• Vulnerability Exploitation
• AI Safety Research

Best for: CTO, VP of Engineering/Data, Executive, AI Security Engineer, Director of AI/ML, Policy Maker

Related on AIssential

• Researchers create AI worm that adapts attacks without human input — An AI-powered worm autonomously adapts attacks, lowering hacker costs and posing new internet security threats.
• AI Agents Enable Adaptive Computer Worms — AI agents power adaptive computer worms that generate tailored attacks and propagate autonomously using stolen compute.
• With $1 Cyberattacks on the Rise, Durable Defenses Pay Off — Generative AI accelerates both cyberattacks and defenses, necessitating a shift to foundational security rather than reactive patching.
• AI Worm — The AI worm prototype integrates an LLM for autonomous, decentralized propagation and multi-attack generation, posing a new cyber threat.
• Google detects AI-assisted cyber exploit before mass attack — AI is now actively used by threat actors to develop zero-day exploits, signaling a new era in cybersecurity.
• Anthropic's Mythos AI model sparks fears of turbocharged hacking — Advanced AI models can both rapidly identify software vulnerabilities and generate exploits, posing significant cybersecurity challenges.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by artifical intelligence via Google News.