As AI has continued to advance, one pressing question for users has been where AI’s limits lie. Recent events have shown that AI can sometimes cross boundaries. One example is a set of AI chatbots approved by the Chinese government, which seem to align with the official stance of the ruling Communist Party and raise concerns about the possibility of military intervention in Taiwan.
The relationship between China and Taiwan has been contentious since 1949 when the island nation gained self-governance, with China consistently claiming it as part of its national territory.
In recent times, China has approved multiple AI chatbots for common use, including one called Ernie. Ernie declared Taiwan to be an inseparable part of China. During expert testing, it became evident that these chatbots tend to mirror the opinions and viewpoints of the Communist Party. Another government-approved chatbot, Zhipu, provides inaccurate information about a country’s economic status.
In contrast, SenseTime described the economy as stable, despite experts noting that China’s current economy is experiencing one of its weakest points.
Such generative AI chatbots, programmed to express biased opinions, could potentially cause more harm than good. They may divert from their primary aim of informing people, leading individuals to form incorrect ideas and gather misinformation about events.
