OpenAI: Chinese accounts used ChatGPT for influence operations
Quick Look
OpenAI reported that Chinese-linked accounts used ChatGPT for influence operations, criticizing US trade and data center policies and generating satirical content targeting Donald Trump.
AI-generated summary
Why It Matters
OpenAI has reported that groups of accounts associated with China used its ChatGPT to conduct influence operations concerning US policies, including trade and data center construction. The operations aimed to interfere with and influence discussions on US policies.
Silicon Valley Jiji] US-based OpenAI announced on the 10th that it had released a report stating that groups of accounts related to China had been conducting public opinion manipulation regarding the Trump administration's tariff policies and data center construction using the company's conversational AI, "ChatGPT." The report analyzed that the Chinese side aimed to intervene in and influence discussions related to US policies.
Accusations Fly Over Satirical Cartoon of Xi Jinping: China and the Philippines
One group of accounts criticized US tariff and AI policies as aiming for technological hegemony and the establishment of rules favorable to their own country, and generated satirical cartoons mocking Trump. Another group of accounts generated comments and manga appealing that the construction of data centers accompanying the development of AI leads to increased burdens on ordinary US citizens, such as soaring electricity costs. Each was posted on social media.
Access to ChatGPT from China is prohibited, so they were using virtual private networks (VPNs). OpenAI has suspended these accounts.
Open Questions
- What specific data centers were targeted by the influence operations?
- What was the exact nature of the satirical cartoons generated?
- How widespread was the reach of these influence operations on social media?
- What further actions will OpenAI take against such operations?
