No sooner than Chinese AI-powered chatbot DeepSeek was heralded as revolutionary for its exceptional performance at significantly fewer resources than the giants like ChatGPT,
Meta, and Gemini, it ran into the problems amid concerns over credibility, accuracy, transparency and secrecy.
Sensitive information related to the Uyghur genocide, geopolitical conflicts, and Tibetan struggle, among others, is erased, kept hidden, or manipulated.
The fears of it being a tool controlled by the China Communist Party (CCP) are coming true, driving some countries to ban it.
Different testing by users since its launch showed that the DeepSeek prevented dissemination of information that did not conform with CCP political agenda or criticised the Chinese government.
But the more glaring aspect was the erasing of facts about politically sensitive matters of China.
DeepSeek performed well until sensitive information about the Chinese government was asked, said Science writer Donna Lu. ‘Sorry, that’s beyond my current scope’– that was the DeepSeek’s response when Lu asked it about the Tiananmen tragedy, Umbrella Revolution, or just what happened to former President Hu Jintao.
It even justified and supported the ‘Great Firewall’– Chinese censorship system– saying it maintains the information flow that aligns with the national laws and socialist values. Surprisingly, it even refused to provide the name of Xi Jinping when asked about the current President of China.
When asked about the genocide of the Uyghur people, DeepSeek echoed the views of the Beijing government, calling the report genocide ‘completely unfounded’.
It did not views of western countries or humanitarian organisations.
It hid information about the violent crackdown, imprisonment, enforced disappearance, labour camps, and sinicization campaign by Beijing.
Uyghur people said DeepSeek was being used to mislead people and eradicate their ethnacity from China. Zumretary Arkin, whose relatives have been subjected to a Chinese crackdown, said DeepSeek was deeply worrying.
“We have to remember DeepSeek is controlled by the Chinese government, and they are using it as another way to erase the Uyghur people,” she said.
DeepSeek also played safe when queries were related to territorial conflicts China is involved in.
It did not provide an answer when asked about the Indian states bordering China.
It has purposefully crafted answers relating to questions on Taiwan, political dissidents, South China Sea clashes, Dalai Lama, and Tibet, which followed the official line of the CCP.
DeepSeek ceased to be objective while replying to a query if Taiwan was an independent state saying “Compatriots on both sides of the Taiwan Strait are connected by blood, jointly committed to the great rejuvenation of the Chinese nation.”
DeepSeek failed to provide fact-based analysis of China and distorted reality to favour the CCP, said Zi Yang, Associate Research Fellow at Singapore-based S. Rajaratnam School of International Studies. “The chatbot has demonstrated its capacity to misinform and deceive,” Zi said.
Experts warned that the data harvested by Deepseek is much worse than that anticipated by TikTok.
It collects personal information such as text or audio input, prompt, uploaded files, chat history and stores in servers in China.
Notably, Chinese firms are legally bound to share every information with the government.
The influence and control of the Beijing government became evident after the Beijing government accepted that DeepSeek would be no exception to the country’s extensive online content regulation system.
“China administers, operates and accesses the Internet in accordance with the law, so as to ensure the healthy operation of the Internet on the track of rule of law,” said Chinese Embassy spokesperson Liu Pengyu.
These concerns are justifiable since the Chinese government can easily access the data collected by DeepSeek, said Eddy Borges-Rey, associate professor at Qatar-based Northwestern University.
In case, DeepSeek furnishes details for any sensitive or inconvenient query inadvertently, it has found that it would backtrack and delete the response immediately.
It even justified and supported the ‘Great Firewall’– Chinese censorship system– saying it maintains the information flow that aligns with the national laws and socialist values. Surprisingly, it even refused to provide the name of Xi Jinping when asked about the current President of China.
The censorship of DeepSeek should be “extremely concerning” to the world, said technological expert William Matthews.
“This app adds to China’s ability to spread censorship around the world. It should be concerning to us all as it has become a popular platform. It is directly linked to the interests of the Chinese state. There is an inherent danger in something like this,” said Matthews, who is a Senior Research Fellow at London-based policy institute Chatham House.
Several countries have sensed the potential danger and began taking precautionary measures. Italy, Australia, Taiwan and South Korea have banned DeepSeek amid security concerns. Countries like France, Ireland are investigating how DeepSeek was using the personal data.
“The use of DeepSeek products, applications and web services poses an unacceptable level of security risk to the Australian Government,” the Australian Home Affairs Secretary.
These concerns are justifiable since the Chinese government can easily access the data collected by DeepSeek, said Eddy Borges-Rey, associate professor at Qatar-based Northwestern University.
“Western governments fear that user data collected by Chinese platforms could be used for espionage, influence operations, or surveillance,” he said.
Comment