Artificial intelligence company Anthropic said it has uncovered and disrupted a large influence operation linked to officials in the United Arab Emirates that allegedly targeted critics of Abu Dhabi’s role in Sudan’s civil war.
The findings were disclosed in a report released on Thursday that examined how the company’s AI systems were being used by state-linked actors and influence networks. Anthropic said it identified the campaign with “high confidence” and took action to stop the activity.
According to the report, individuals connected to the operation used Anthropic’s Claude chatbot to create reports, advocacy documents, social media content, and testimony related to the conflict in Sudan.
The campaign was primarily focused on the Muslim Brotherhood, a political and religious movement that has long been viewed as an opponent by several governments across the Middle East.
Anthropic said the operation extended beyond political messaging and included efforts to influence discussions surrounding Sudan’s ongoing conflict.
Sudan has been engulfed in war since April 2023, when fighting erupted between the Sudanese Armed Forces and the paramilitary Rapid Support Forces. The conflict has devastated large parts of the country and displaced millions of people.
The United Nations has repeatedly described the crisis as one of the world’s worst humanitarian emergencies, with widespread displacement, food insecurity, and growing concerns over civilian protection.
The war has also become the subject of international political disputes.
The UAE has faced accusations from Sudanese authorities and United Nations experts that it has supported the Rapid Support Forces during the conflict. Critics argue that such support has contributed to the continuation of the war.
Abu Dhabi has consistently denied those allegations and rejected claims that it has fueled the conflict.
Anthropic’s report stated that one part of the influence operation involved compiling what the company called “counter-accountability dossiers” on United Nations special rapporteurs and experts who had criticized the UAE’s alleged conduct in Sudan.
According to the company, the materials appeared designed to challenge or undermine the credibility of international investigators and human rights officials.
The report also described another effort that allegedly involved using the identity of a Sudanese human rights organization.
Anthropic said individuals behind the operation drafted testimony intended for submission to the United Nations while presenting the material as if it originated from independent local witnesses.
The company stated that the objective appeared to be ensuring that messaging favorable to one side of the Sudan conflict would reach international institutions under the appearance of neutral civil society testimony.
Anthropic named two individuals whose identities were allegedly referenced during the effort to prepare testimony materials.
The company said its investigation found evidence that AI-generated content was used to support broader influence activities across multiple platforms and formats.
The case highlights growing concerns about the use of advanced artificial intelligence tools in political influence campaigns.
As AI systems become more capable, technology companies have increased efforts to monitor misuse involving misinformation, disinformation, covert influence operations, and deceptive online activities.
Anthropic regularly publishes reports detailing attempts by state-linked groups, criminal organizations, and other actors to use its technology for activities that violate company policies.
The latest findings add to broader international concerns about how artificial intelligence can be used to shape public opinion, influence institutions, and affect political debates.
Technology companies have warned that AI tools can rapidly generate large volumes of persuasive text, making it easier for influence operations to produce reports, social media posts, public statements, and advocacy materials at scale.
Researchers say such capabilities can amplify existing political campaigns and complicate efforts to identify the true origin of information.
Anthropic did not disclose how long the alleged operation had been active or whether similar campaigns were identified on other AI platforms.
The company said it removed access connected to the activity after completing its investigation.
The report comes as governments, technology firms, and international organizations continue debating how to regulate artificial intelligence and prevent misuse while preserving the benefits of rapidly advancing AI technologies.
Anthropic’s findings are likely to intensify those discussions, particularly as concerns grow about the role AI could play in future geopolitical disputes, information campaigns, and international conflicts.
