Beijing Monopolizes Global AI Knowledge Interpretation, Threatening Taiwan Culture and Security

In an era of rapid artificial intelligence development and algorithms deeply intervening in human daily thought, language and data are no longer merely tools of communication, but have become core battlefields for competing for discourse power and ideological influence. Hung Tzu-han, a research fellow at Academia Sinica's Institute of European and American Studies, recently issued a profound warning, pointing out that the Beijing authorities are attempting to completely monopolize the right to interpret knowledge in the global Chinese-speaking world on the training foundation of artificial intelligence and big data by massively supplying Simplified Chinese digital data. This strategy is not merely a technological and cultural export; through modern cognitive warfare tactics, it also inflicts a new type of epistemic injustice on Chinese-speaking users outside mainland China. For Taiwan, which has long used Traditional Chinese and emphasizes diverse democratic values, this warning concerns the survival of national security and cultural subjectivity.

Reviewing the current context of global artificial intelligence development, the training of large language models and generative artificial intelligence relies heavily on massive amounts of textual data. Since the vast majority of Chinese-language data available on the internet for machine learning consists of content based on mainland China's official standards and Simplified Chinese, artificial intelligence often unconsciously replicates and internalizes Beijing's official narrative and ideology when generating responses, interpreting historical events, or defining political concepts. Hung Tzu-han pointed out that the Beijing authorities are not passively letting data flow onto the internet, but are intentionally and systematically instilling Simplified Chinese data with specific perspectives into the world through various digital channels and policy tools. This phenomenon causes global Chinese-speaking communities to frequently fall into the dilemma of "being represented" when accessing information generated by artificial intelligence, which is tantamount to depriving them of the possibility of diverse interpretations.

This phenomenon caused by data monopoly is academically described as a new form of epistemic injustice. Traditional epistemic injustice mostly focuses on the marginalization of social groups in the knowledge production process; however, in the digital and algorithmic era, when artificial intelligence becomes the primary intermediary for most people to acquire knowledge and viewpoints, the power asymmetry behind the data has become the greatest hidden concern. Mainland China has long used the Great Firewall to filter internet information and strongly dominated public opinion and historical interpretation through the state apparatus. When these Simplified Chinese data, carrying strong political censorship and a single perspective, are heavily fed into globalized algorithmic models, the algorithms treat them as the "norm" and "truth" of the Chinese-speaking world. This causes the diverse experiences, historical memories, and democratic values possessed by Taiwan, Hong Kong, and other overseas democratic Chinese-speaking communities to be systematically ignored and marginalized in the underlying logic of algorithms.

Facing this silent yet far-reaching cognitive and cultural campaign, scholars have clearly suggested that the Taiwanese government must re-examine and adjust its domestic language and cultural policy. The primary task is to gradually abolish, in accordance with the law, the policy of exclusively exalting Mandarin Chinese, which previously carried a single cultural hegemony. As a major hub of the global Traditional Chinese-speaking world, Taiwan has not only completely preserved the aesthetics and cultural heritage of Chinese characters, but has also developed a unique linguistic context rooted in freedom, democracy, and a pluralistic society. Hung Tzu-wei believes that facing up to and gradually replacing traditional single-Chinese-language thinking with Taiwan's native languages and diverse Taiwanese writing systems will be an inevitable trend for Taiwan to establish subjectivity in the cultural and digital fields and counter cognitive warfare in the future. This is not only a shift in language policy, but also an important line of defense to establish national cultural identity and ensure knowledge autonomy.

Furthermore, the issues involved in this topic have long surpassed mere academic discussion; instead, they constitute a strategic defense issue that Taiwan must seriously face when confronting the digital expansion of a major power. As artificial intelligence technology gradually penetrates public policies, educational systems, and even evaluation models for national security defense, the diversity and autonomy of data sources become crucial. If Taiwan cannot establish its own knowledge production and interpretation system in the digital world, and allows Simplified Chinese data and Beijing's official narrative to completely occupy the commanding heights of algorithms, Taiwan's social cognition and democratic values may be imperceptibly eroded. Therefore, all sectors of industry, government, academia, and research must face up to this struggle for the right of interpretation triggered by AI and big data, and actively construct digital databases and language application environments possessing Taiwanese subjectivity, so as to stand firm in the globalized wave of intelligence.

(Source: Central News Agency)