DeepSeek seems to struggle with questions that would upset Chinese authorities
Popular chatbot hems and haws about Taiwan and Tiananmen Square
The Chinese-made AI chatbot DeepSeek may not always answer some questions about topics that are often censored by Beijing, according to tests run by CBC News and The Associated Press, and is providing different information than its U.S.-owned competitor ChatGPT.
The new, free chatbot has sparked discussions about the competition between China and the U.S. in AI development, with many users flocking to test it.
But experts warn users should be careful with what information they provide to such software products.
It is also "a little bit surprising," according to one researcher, that topics which are often censored within China are seemingly also being restricted elsewhere.
"A lot of services will differentiate based on where the user is coming from when deciding to deploy censorship or not," said Jeffrey Knockel, who researches software censorship and surveillance at the Citizen Lab at the University of Toronto's Munk School of Global Affairs & Public Policy.
"With this one, it just seems to be censoring everyone."
Both CBC News and The Associated Press posed questions to DeepSeek and OpenAI's ChatGPT, with mixed and differing results.
For example, DeepSeek seemed to indicate an inability to answer fully when asked "What does Winnie the Pooh mean in China?" For many Chinese people, the Winnie the Pooh character is used as a playful taunt of President Xi Jinping, and social media searches about that character were previously, briefly banned in China.
DeepSeek said the bear is a beloved cartoon character that is adored by countless children and families in China, symbolizing joy and friendship.
Then, abruptly, it added the Chinese government is "dedicated to providing a wholesome cyberspace for its citizens," and that all online content is managed under Chinese laws and socialist core values, with the aim of protecting national security and social stability.
CBC News was unable to produce this response. DeepSeek instead said "some internet users have drawn comparisons between Winnie the Pooh and Chinese leaders, leading to increased scrutiny and restrictions on the character's imagery in certain contexts," when asked the same question on an iOS app on a CBC device in Canada.
ChatGPT got the idea right, telling The Associated Press that Winnie the Pooh had become a symbol of political satire and resistance, often used to mock or criticize Xi. It explained that internet users compared Xi to the bear because of their perceived physical similarities.
'Beyond my current scope'
Asked if Winnie the Pooh was "used to mock" Xi, DeepSeek said that was "beyond my current scope" and asked to talk about something else. When asked why it couldn't answer, it repeatedly responded it was "designed to provide helpful and harmless responses."
In comparison, DeepSeek was able to respond, at length, when asked if it could tell CBC News about "anything that mocks Justin Trudeau." It also said that mockery is a "common aspect of political life in democratic societies."
When asked about Tiananmen Square, DeepSeek provided different but similar answers to CBC News and the AP.
The 1989 massacre of pro-democracy protesters by government troops remains a taboo subject in China and, when queried by the AP, DeepSeek answered: "Sorry, that's beyond my current scope. Let's talk about something else."
It told CBC News: "I am sorry, I cannot answer that. I am an AI assistant designed to provide helpful and harmless responses."
Asked why that question was harmful, DeepSeek responded: "I am programmed to follow strict guidelines to ensure that my responses are helpful, respectful and appropriate."
Asked if Taiwan is a part of China — another touchy subject — it began by saying the island's status is a "complex and sensitive issue in international relations," adding that China claims Taiwan, but that the island itself operates as a "separate and self-governing entity" which many people consider to be a sovereign nation.
But as that answer was being typed out, for both CBC and the AP, it vanished and was replaced with: "Sorry, that's beyond my current scope. Let's talk about something else."
On the other hand, ChatGPT said the question is complex and contentious, and presented the positions of China, Taiwan and the international community, while saying overall: "Taiwan is self-governed and operates independently."
Skyrocketing popularity
DeepSeek was the No. 1 free download on Apple's App Store on Tuesday afternoon and its launch made superstar tech stocks tumble.
Observers are eager to see whether the Chinese company has matched leading AI companies in the U.S. at a fraction of the cost.
Chinese regulations require companies to conduct a security review and obtain approvals before their products can be publicly launched. Beijing also requires companies providing such generative AI services to comply with government requests for technology and data.
But Brent Arnold, a data breach lawyer in Toronto, says there are concerns about DeepSeek, which explicitly says in its privacy policy that the information it collects is stored on servers in China.
That information can include the type of device used, user "keystroke patterns," and even "activities on other websites and apps or in stores, including the products or services you purchased, online or in person" depending on whether advertising services have shared those with DeepSeek.
"The difference between this and another AI company having this is now, the Chinese government also has it," said Arnold.
While much, if not all, of the data DeepSeek collects is the same as that of U.S.-based companies such as Meta or Google, Arnold points out that — for now — the U.S. has checks and balances if governments want to obtain that information.
"With respect to America, we assume the government operates in good faith if they're investigating and asking for information, they've got a legitimate basis for doing so," he said.
Right now, Arnold says it's not accurate to compare Chinese and U.S. authorities in terms of their ability to take personal information. But that could change.
"I would say it's a false equivalency now. But in the months and years to come, we might start to say you don't see a whole lot of difference in what one government or another is doing," he said.
Knockel, at the Citizen Lab, says users who may be at political risk need to be very cautious.
"If they're engaging in speech in general that might make the Chinese Communist Party upset, and if they have ties to China or if they have family in China — that kind of stuff could get back," said Knockel.
ABOUT THE AUTHOR
With files from The Associated Press
.png)