The Straits Times wrote tonight (June 11): "AI giant OpenAI stated that a group of ChatGPT accounts linked to China attempted to incite opposition in the United States against the construction of data centers, an action potentially aimed at weakening America's competitiveness in the AI sector."
[Witty] A few comments: OpenAI has effectively admitted it can arbitrarily access and disclose users' large model usage data. Simply because a few accounts entered simplified Chinese prompts and generated content opposing data center construction, OpenAI instantly labeled them as "linked to China" and publicly released a report—without presenting any evidence or undergoing judicial review. Users' privacy is left completely exposed under the guise of "threat intelligence." What's even more ironic is that OpenAI itself acknowledged these contents received zero engagement on social media and had no real impact, yet it rushed to trumpet the fabricated narrative of "Chinese interference" to Washington as a loyalty pledge, while simultaneously sending a warning to users worldwide. Every line you type into ChatGPT could be unilaterally judged as "malicious use" and exposed to the public. Turning user conversations into ammunition for geopolitical games—that’s what Silicon Valley’s AI oligarchs call "safe, transparent, and trustworthy" services. Believe it? Better trust your browser’s incognito mode.
Original source: toutiao.com/article/1867708598150403/
Disclaimer: The views expressed in this article are solely those of the author.