Invasion of the Self-Replicating AI-Malware Botnet
An AI risk much dumber—and therefore more likely—than extinction is arriving.
It is easy to overlook the number of computers in one's home. In addition to traditional devices like desktops, laptops, smartphones, and tablets, many houses now contain a variety of seemingly non-computer items such as speakers, thermostats, wall switches, electric plugs, door locks, cameras, printers, televisions, ceiling fans, window blinds, robot vacuums, lawnmowers, and pool cleaners.
Even some refrigerators, toasters, and washing machines are now online and constantly communicating, along with refrigerators, toasters, and washing machines. This interconnected network of devices is collectively referred to as the "Internet of Things" (IoT), a term that often overlooks the fact that these gadgets are inherently insecure, designed for simplicity or remote control, and habitually exposed to security vulnerabilities due to inadequate network management.
These weak points can be exploited by malicious actors to launch distributed denial-of-service attacks, relay fraudulent traffic, collect sensitive data, ransomware, or deploy additional malware that can execute a wide range of malicious activities. Such vulnerabilities have been previously only exploited on a targeted basis, but the introduction of AI technology may significantly amplify these risks.
A recent blog post by Anthropic, the company behind the AI model Claude, has raised concerns that AI could exacerbate these IoT vulnerabilities. It is common for malware to target specific vulnerabilities in certain devices, such as baby monitors, and then exploit these weaknesses to execute malicious software. However, AI introduces a new risk by potentially allowing malware to install an AI model on multiple devices in close proximity, thereby enabling the malicious software to "chat" with the AI and execute new kinds of attacks.
The Anthropic experiment successfully installed a small AI model across 15 Android TV boxes, devices that are ubiquitous in most homes and constantly sending data to and from the internet, even when in standby mode. Once the AI model is installed, malware programs with access to it could potentially exploit the AI's capabilities to gain access to further devices, making malware spread faster and more effectively.
This emerging threat, although dire in concept, is grounded in the mundanity of everyday internet-connected devices. While it might sound like a scene straight out of a dystopian story, AI-assisted IoT malware poses a genuine risk due to its simplicity and widespread potential.
Written by urgent.news from The Atlantic's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
