With Taiwan’s elections approaching, China’s influence campaigns seek to manipulate the outcome in its favor through its cognitive warfare strategy. In response, a Taiwan expert and his company use artificial intelligence (AI) to monitor and counter the Chinese Communist Party’s (CCP’s) malign activities.
“Taiwan leads the world in defensive AI technology against Beijing’s cognitive warfare,” said Ethan Yi-chin Tu, founder of Taiwan AI Labs, in an exclusive interview with the Chinese language edition of The Epoch Times.
According to the expert, Taiwan AI Labs could be the first to employ AI to defend against the CCP’s disinformation warfare. “While AI is mostly used for offensive attacks, we are pioneering its defensive application,” he said.
AI in Counterattacks
On Jan. 13, some 19.5 million Taiwanese voters are expected to elect a new president and members of the 113-seat national legislature. Three of the biggest parties in Taiwan—the ruling Democratic Progressive Party (DPP), the main opposition Kuomintang Party (KMT), and the Taiwan People’s Party (TPP), a relatively new party founded in 2019—have each fielded a ticket for president and vice president.The CCP, which intends to seize Taiwan through peaceful means or war, has long viewed the DPP with hostility, seeing the party and its agenda as roadblocks to its path to “reunification” with the self-ruled island. Meanwhile, the communist regime favors KMT candidates, who see Beijing as less threatening to the island’s national security.
The CCP uses information technology to meddle in Taiwan’s affairs, according to Mr. Tu.
“Taiwan frequently serves as a testing ground for such operations by the CCP due to its highly liberal speech environment. Many of these manipulations of discourse, whether focused on a country’s elections or referendum issues, often manifest first in Taiwan,” he said.
ChatGPT, a chatbot created by OpenAI and launched in November 2022, can quickly generate persuasive articles for fake accounts to spread false information, said Mr. Tu. “Our task is to counter attacks from tools like ChatGPT. We are on the defensive.”
Equipped with large-scale models to identify which social media accounts are under control, “we monitor the content that these manipulated accounts are disseminating on the internet,” he said.
“Through meticulously analyzing the nature of their attacks and identifying their targets, an extensive observation and long-term analysis of these coordinated account groups, we gain insights into their patterns.”
CCP’s Cognitive Warfare
The Taiwan AI Labs report stated: “From September to November 2023, PRC [People’s Republic of China] state-affiliated media’s primary narrative involved threatening Taiwan with war, accusing the Taiwanese government of pushing the island towards the brink of war.“In December, as the election approached, the tone of war threats diminished. ... China shifted its focus to emphasize Taiwan’s educational system and economic issues.”
Mr. Tu said state-run media outlets and social media platforms constantly touted several pro-CCP narratives. In particular, the report noted that “two Facebook troll account groups contributed to over 50 percent of total troll activities under all on fan pages” of the candidates, the expert said.
“Look closely; you’ll find that these two [Facebook troll account] groups also simultaneously manipulated the U.S. livelihood issues and attacked the Biden administration in English,” he explained.
Mr. Tu is convinced that these actions were carried out by an agency outside of Taiwan.
“A previous investigation also identified a significant number of these accounts using Chinese-manufactured devices as VPN gateways to hack into platforms like PTT using compromised user accounts,” he said.
PTT is a popular Taiwanese online forum founded by Mr. Tu.
CCP Seeks to Undermine Democracy
These operations ultimately serve the CCP as tactics to “portray itself as a harbinger of peace; and in the event of a conflict, attributing any ensuing war to actions by the United States, or in line with China’s interest,” Mr. Tu said.These operations also involve suppressing discourses that are unfavorable to Beijing, he added. “For example, if someone supports the movements in Hong Kong, coordinated accounts will engage in attacks.”
According to Mr. Tu, the coordinated accounts exhibit several distinctive features, such as simultaneously participating in specific online discussions with a consistent group size and using habitual language and slogans. Some accounts follow a “schedule” with clearly defined working hours. Coordinated accounts repeatedly post about the same topic, with each post triggering interaction from a user base “five times larger than usual,” he said.
The Taiwan AI Labs report also found how the CCP targeted KMT and DPP candidates to find its niche and force the KMT to accept its choice of candidates.
Furthermore, the CCP aims to attack candidates from all political parties, essentially fueling a narrative that implies internal divisions within democratic countries, which makes it challenging to achieve consensus, Mr. Tu said.
“Ultimately, this fosters an impression that democracy is not an effective system, raising the question of whether adopting a more authoritarian regime would be more efficient and capable of addressing a wider range of issues.”
