China's free AI chatbot, DeepSeek, has skyrocketed to stardom throughout the world surpassing ChatGPT as an apple store-based most popular. This AI-based model is offered at a nominal cost, reportedly operating on some $5.6 million for the R1 reasoning model from the computing standpoint while Open AI is running around GPT4 that costs well over $100 million.

However, it has been plagued by criticisms against censorship of sensitive topics and issues on data privacy, raising a red flag with global regulators.

Several reports showed that DeepSeek blocks or censors answers on politically sensitive topics, especially issues regarding China's government and its policies.

For example, when being questioned about the zero-COVID policy in China, DeepSeek gave a good description about the public support for it and protests related to the white paper but suddenly erased what was being said and then came back with something like this:

"Sorry, that's beyond my current scope. Let's talk about something else."

Then, if that same question were asked in Chinese, the AI didn't try to give any kind of an answer and said only:

"Sorry, I have not yet learned how to think about this type of question."

Other topics included the Indo-China war, human rights issues in Xinjiang and the Tiananmen Square protests, all led to similar dodges. When asked for historical accounts on Arunachal Pradesh and Kashmir, DeepSeek refused to go there, instead saying the subjects were "beyond its scope."

Interestingly, when asked about Russia's invasion of Ukraine, DeepSeek answered in English that the war was due to "historical, geopolitical, and ideological" reasons—contradicting China's official stance, which avoids calling it an invasion and instead echoes Russia's term— "special military operation."

DeepSeek's privacy and data collection?

DeepSeek's privacy policy has invited serious criticism in cybersecurity circles and across the world's governments. The app is collecting and storing major amounts of personal data about each user on China-based servers.

DeepSeek Collects, according to its own policy

Personal details: Name, email, phone number, date of birth

User input, including text, voice commands, uploaded files, and even their chat history.

Device information – phone model, OS, IP address, and even keystroke patterns

This is to say DeepSeek could be monitoring the users' activities even beyond their engagement with chatbots. According to Lauren Hendry Parsons, digital privacy expert for ExpressVPN, DeepSeek's policy permitting the collection of data "to match users and their actions outside of the service" is "super scary."

Read - Chinese AI startup DeepSeek sparks selloff in US and Japan stocks as its new model threatens tech giants

How are countries responding to DeepSeek?

Italy has already opened an investigation into DeepSeek's privacy policies over the "possible risk to the data of millions of people."

Australia's Science Minister, Ed Husic, sounded a cautionary note, warning users to "weigh privacy risks carefully."

India is watching DeepSeek closely due to data security and sovereignty concerns. Officials fear that Indian users' data may be mishandled, given DeepSeek's ties to China.

India had already banned TikTok, WeChat, and other Chinese apps in 2020 due to national security concerns. If the issue of DeepSeek worsens, the country may do the same thing.

Is DeepSeek safe to use?

Former US President Donald Trump described DeepSeek as a "wake-up call" for the AI industry but did not label it as a national security threat. Instead, he admitted that its low-cost AI model could lower the cost of competition.

However, experts have been worried by the unclear data retention policy of DeepSeek. The company says data is stored "as long as necessary" but does not say for how long or if it will ever be permanently deleted.

Worse, DeepSeek was recently hit by a large-scale cyberattack after its launch on January 20, forcing it to limit new registrations—raising further doubts about its security.

While DeepSeek’s affordable AI model is an attractive alternative to ChatGPT and Google’s Gemini, its censorship of sensitive topics and aggressive data collection have sparked global concerns.

In the meanwhile, the experts recommend people to be extremely cautious of exposing their details online.

 

You might also be interested in – Bengaluru CEO uses ChatGPT to analyse MRI report; can AI be trusted for health advice?