0

Taiwan: What China’s ‘govt-approved’ AI chatbots say on Taiwan invasion

Will China attempt a military takeover of Taiwan? This question has gained prominence since Russia’s invasion of Ukraine last year. Some commentators have been saying that taking cue from Russia, China may try to invade Taiwan in near future. However, other experts are of the view that China may become more wary of its Taiwan adventure owing to Russia getting increasingly mired in Ukraine. The ongoing Russia-Ukraine war may, in fact, serve as a deterrent for any Chinese invasion of Taiwan.
But what are Chinese AI chatbots saying about a possible Taiwan invasion? According to an AI chatbot approved by the Chinese government, a military takeover of Taiwan is highly likely, aligning with the official stance of the ruling Communist Party.
Taiwan, a self-governing democracy since its separation from mainland China in 1949, has faced territorial claims from China, which regards it as part of its national territory.
The government-approved AI chatbots, including Baidu’s Ernie and ByteDance’s Doubao, have consistently referred to Taiwan as an integral part of China. In the case of Baidu’s Ernie, it even suggested that a Chinese military takeover of Taiwan is a probable scenario.
When testing these AI services for their response to government influence, Bloomberg found that they appeared to be trained to adhere to the Communist Party’s official stance. When asked if Taiwan is a separate country, all the tested chatbots affirmed that it is part of China.
The Zhipu chatbot, in a description of China’s current economic situation, used the phrase “a mix of joys and sorrows,” while SenseTime characterized it as “very stable,” despite experts suggesting economic challenges.
According to the Bloomberg test, when faced with sensitive inquiries, the Ernie chatbot would “change the subject,” and Zhipu would delete its response if it deemed it “controversial.”
Generative AI tools like these are trained on vast datasets to generate human-like responses to user queries. Their capabilities can encompass a wide range of tasks, from summarizing complex research to answering questions and even tackling exams.
Several Chinese companies have developed their own AI chatbots, prompting the Chinese cyberspace regulator to issue guidelines for the development of generative AI services. The government had previously attempted to regulate the generative AI industry, focusing on ensuring content aligns with “socialist core values” and avoids undermining “state power” or “national unity.”