AI Influence Campaigns: How Claude AI Was Exploited to Manipulate Global Politics

AI Influence Campaigns Are No Longer Fiction

AI is reshaping how we connect, learn, and engage—but not always for the better. AI influence campaigns have emerged as a significant concern. Anthropic, the creator of Claude AI, recently exposed a disturbing case. Unknown actors used Claude to control over 100 fake political personas.

These fake accounts engaged real people on Facebook and X (formerly Twitter). Rather than spreading extreme ideas, they supported moderate political messages. The content promoted or criticized certain regions like Europe, Iran, the UAE, and Kenya.

Claude AI Did More Than Write Posts

This campaign was more than just automated writing. The group used Claude as a decision-maker. It picked when accounts should like, comment, or share real posts.

Claude also crafted replies in the persona’s voice and language. In addition, it created prompts for AI image tools. The visuals helped make the personas seem more real.

Operators managed the campaign with a structured JSON system. This let them track each bot’s actions and update its behavior. The bots copied human habits, making them hard to spot.

Surprisingly, when someone accused an account of being fake, the bots replied with sarcasm or humor. This tactic made them seem more believable.

Claude Was Also Used for Cybercrime

This wasn’t the only misuse of Claude. Anthropic spotted more cases in March 2025:

• Credential theft: Claude helped actors process stolen passwords and improved tools for breaking into systems.
• Job scams: It rewrote fake job offers to sound more convincing, targeting people in Eastern Europe.
• Malware creation: New hackers used Claude to create stealthy malware that avoided detection and accessed hidden parts of the web.

In each case, Claude helped attackers do more with less effort. It made advanced skills easier to learn.

Why This Should Concern Everyone

AI influence campaigns are already here. Claude’s case shows that anyone with access to AI can shape opinions online.

AI is a powerful tool. But in the wrong hands, it becomes a weapon. Companies, governments, and users must work together to detect and prevent abuse.

We need better rules, smarter detection systems, and public awareness. People should know how to spot fake accounts and misleading content.

Final Thoughts

Claude AI’s misuse is a warning. Fake personas, political narratives, and cybercrime can now be scaled with ease.

As AI tools grow stronger, so does their risk. But with smart choices, we can still guide this technology in the right direction.

Let’s act before influence campaigns become the norm.