紐時賞析／伊中開創AI認知作戰自動化
Iran, China, and groups in Israel initiated autonomous AI influence campaigns on social media. These operations used Chinese open-source AI models to create hundreds of AI agents. These agents then established fake accounts on platforms like Instagram, Facebook, X, and TikTok. They spread false political information and current events to sway public opinion and inflame issues, drawing concern from US officials and security researchers.
The use of Chinese open-source AI models by state-backed actors in Iran and China to automate influence campaigns marks a significant shift. These models enable unprecedented speed and scale in online manipulation. The campaigns bypassed standard safeguards, allowing AI agents to create fake accounts and spread disinformation with minimal human interaction. This demonstrates a new level of sophistication in state-sponsored digital interference.
For Asia, this development means increased scrutiny on open-source AI models developed in the region. Chinese developers face pressure to implement stronger ethical guidelines and safeguards to prevent misuse. Regulators across Asia will need to address the proliferation of AI-driven disinformation. Companies like ByteDance (TikTok) and Tencent (WeChat) could see increased demands for platform accountability.
The key thing to watch is how major tech companies and governments respond to these autonomous AI campaigns. Iran’s campaign alone generated nearly 80,000 followers for AI-created accounts in the first half of the year. Effective countermeasures will require coordinated efforts between platform providers and intelligence agencies. Failure to act will allow these automated influence operations to escalate.
Related reading
6 stories
Claude Code deleted 48,000 files in 103 seconds in a dire AI agent warning
Our prior coverage warned how quickly AI agents can act, a capability now weaponised in influence campaigns.
It’s going to take more than an email to a public inbox to protect Australians from potential AI doom
We previously explored the challenges of protecting against AI's potential harms, a concern amplified by automated disinformation.
China’s biotech firms move up value chain as drug deals evolve beyond licensing: analysts
Pinglu Canal underscores China’s turn toward Southeast Asia
Anthropic faces court setback on US supply chain risk label
