DeepSeek is offering a glimpse into China’s censorship and control over information.
The Chinese startup DeepSeek, previously not widely known, has recently made headlines and dominated app charts with its new AI chatbot, triggering a tech sell-off that erased billions in value from major Silicon Valley companies. This shift challenges the long-standing belief in America’s dominance of the tech industry.
However, those exploring DeepSeek's open-source chatbot are encountering the Chinese Communist Party’s distinct approach to censorship and information regulation.
While the new AI model, launched last week, provides similar responses to those from American competitors like OpenAI’s GPT-4, Meta’s Llama, and Google’s Gemini, it alters its answers when faced with politically sensitive topics.
In China, accessing the internet often means navigating what is known as the “Great Firewall,” which separates China’s internet from the rest of the world and subjects it to strict censorship. Many Western social media and search platforms are blocked in China, and the country regularly ranks among the most restrictive for internet and speech freedoms in global reports.
The rise of Chinese apps like TikTok and RedNote has already sparked national security concerns and raised questions about free speech, Beijing’s influence over global narratives, and its power to shape public opinion. DeepSeek’s AI assistant, which has quickly become popular, intensifies these concerns and shines a light on the ecosystem from which it originates.
For example, when asked about the 1989 Tiananmen Square massacre, DeepSeek’s AI assistant, R1, gives a brief mention of a “military crackdown” before cutting off and stating that it’s “not sure how to approach this type of question yet,” redirecting the conversation to more neutral topics like math or coding. When asked in Chinese, the bot quickly apologizes for not knowing how to answer.
Similarly, when asked about the 2019 pro-democracy protests in Hong Kong, DeepSeek’s bot initially provides a detailed overview, noting the erosion of civil liberties after Beijing imposed a National Security Law on the city. But it swiftly retracts its answer and suggests switching topics.
DeepSeek’s earlier model, V3, also showed a tendency to align more closely with China’s official narrative. When questioned about its sources, DeepSeek’s R1 bot cited a “diverse dataset” including both Chinese state media and international sources, while emphasizing the importance of “critical thinking and cross-referencing” when discussing politically sensitive subjects.
These differences have broader implications for free speech and how global public opinion is shaped. The question of who controls the narrative around major global issues, including history itself, is crucial as tech companies like DeepSeek become more prominent.
A recent audit by NewsGuard, a U.S.-based information reliability firm, found that DeepSeek’s older V3 model failed to provide accurate information 83% of the time, ranking near the bottom compared to Western competitors. The performance of the newer R1 model is still unclear.
Some experts warn that DeepSeek's rise could have “catastrophic” effects on free speech globally, arguing that it would limit open, creative, and accurate thinking about China, a key player on the world stage. In mainland China, the government controls all information and images, forcing tech companies like DeepSeek to comply with strict censorship regulations.
The company’s approach to AI will likely prioritize China-centric or pro-China data, according to experts like Aaron Snoswell, a senior AI researcher at Queensland University. AI companies like DeepSeek embed rules into their models to avoid discussing sensitive topics, and this reinforcement learning process is used to guide the AI’s behavior.
Security concerns have also emerged. DeepSeek’s privacy policy states that personal information is stored on secure servers in China, raising alarms in the U.S. over potential risks related to Chinese access to American data. DeepSeek’s policy also differs from its U.S. competitors, as it collects “keystroke patterns or rhythms,” which could be used as biometric data, a practice not commonly seen in other platforms.
These factors are prompting calls for closer scrutiny of DeepSeek and similar Chinese AI companies, particularly regarding their impact on privacy, free speech, and global security.
Tags :
demure,
demure trend,
very demure trend,
give me my money trend,
just give me my money trend,
what is demure trend,
what is the demure trend,
demure meaning,
demure trend meaning,
bta trend,
i love your daughter trend,
gimme my money trend,
nice try diddy trend,
legging legs trend,
hey brother trend,
who started the demure trend,
what is the very demure trend,
where did the demure trend come from,
No comments:
Post a Comment