Fake accounts drove the DeepSeek AI hype and distorted markets
12 days ago
- #bot-networks
- #disinformation
- #AI-hype
- DeepSeek, a Chinese AI model, gained rapid popularity, but research shows much of the hype was artificially generated by fake accounts.
- Analysis of 41,864 profiles revealed 3,388 fake accounts (15% of engagement), which coordinated to amplify DeepSeek's visibility.
- Fake accounts used tactics like mutual amplification, piggybacking on trending posts, and synchronized timing to create artificial virality.
- These accounts exhibited hallmarks of state-linked bot networks, including recent creation dates, avatar recycling, and identical posts.
- The manufactured hype briefly influenced US markets, wiping out billions in valuation and distorting perceptions of China's AI capabilities.
- Organizations are advised to use specialized disinformation detection tools rather than building internal solutions due to complexity and time constraints.
- A 90-day rollout plan for disinformation detection includes monitoring, stress-testing with simulations, and developing response playbooks.
- Platforms like X and TikTok are most exploited due to weaker controls and high virality potential.
- Small firms can also be targeted, as hype tactics are cheap and scalable.