Anthropic reported that an influence operation linked to the United Arab Emirates (UAE) used the Claude AI language model as part of a campaign targeting the Muslim Brotherhood. According to the company, the operation employed roughly 300 fake social-media accounts, created a front non-governmental organization, ghost-wrote testimony intended for use at the United Nations, and compiled dossiers on European lawmakers, journalists and UN officials.
Anthropic said it traced the campaign “with high confidence to UAE government officials.” The UAE Foreign Ministry did not respond to requests for comment after the report was released.
Abu Dhabi has designated the Muslim Brotherhood as a terrorist organization since 2014 and has made countering political Islam a pillar of its regional policy. The United States considers branches of the Brotherhood in Egypt, Jordan, Lebanon and Sudan to be terrorist organizations.
Broader findings on AI misuse
The report highlighted additional troubling examples of how artificial intelligence and related technologies have been misused. Anthropic documented Russian cyberattacks, Chinese surveillance targeting Uyghurs, and Yemeni-developed software connected to armed drones, bombs, missiles and other weapons systems.
Why the report matters
The findings illustrate that advanced language models and other AI tools can be repurposed for influence operations, disinformation campaigns and even to support weapons-related capabilities. The allegations implicate state-linked actors and involve international institutions and officials, underscoring the geopolitical and security implications of AI misuse.
Mohammed Sergie
