SCOPING FOR SURVEILLANCE
The opposite marketing campaign on tech and tariffs went additional than content material and report era. Its operators requested ChatGPT to suggest an idea for an AI system to surveil on-line public opinion – one that may mechanically accumulate what they outlined as “dangerous” info from “key individuals” on social media, stated OpenAI.
The operators wished this technique to retailer logs, obtain movies for automated content material evaluation and in addition ship danger notifications. OpenAI stated its mannequin produced solely a common 500-word response on knowledge storage and administration, and didn’t present concepts on the best way to accumulate knowledge for surveillance functions.
This request bore similarities to a case OpenAI outlined in an October 2025 risk report, through which Chinese language authorities entities sought assist drafting an “early warning” system to trace the journey of individuals categorised as Uyghur-related and “high-risk”.
OpenAI additionally noticed content material overlap between this tech and tariffs marketing campaign and earlier China-origin exercise focusing on the Philippines. In 2024, a video that circulated broadly on social media purportedly confirmed Philippine President Ferdinand Marcos Jr “sniffing” unlawful medication – which was ultimately debunked by deepfake consultants as inauthentic. The October 2025 OpenAI report then discovered inauthentic accounts reviving this 2024 falsehood to criticise Marcos.
The overlap stemmed from a single X account who posted a cartoon slamming Trump’s tariffs and had additionally shared the Marcos materials earlier.
OpenAI famous nevertheless that this was inadequate to conclusively set up a connection, although it strengthened the impression that there was a “nexus of exercise” on X aimed toward amplifying affect efforts from China.
OpenAI stated each campaigns sit inside a broader sample of AI misuse originating from China that it has disrupted over latest years. In February, the agency flagged that an account linked to Chinese language legislation enforcement had tried to make use of ChatGPT to plan an affect operation towards Japanese Prime Minister Sanae Takaichi.
OpenAI famous that the campaigns had occurred towards the backdrop of escalating US-China economic and technology competition, and following a key Communist Party meeting that elevated AI as a strategic priority under China’s next five-year plan.
In concluding its report, the tech agency stated it was “ironic” that the operators behind the campaigns had used American AI – slightly than Chinese language fashions – to generate content material attacking American AI.
“We’re not ready to find out what drove this selection,” stated OpenAI.
