DeepSeek reopens new user sign-ups despite ongoing security concerns

Chinese AI chatbot DeepSeek may have been limiting access to new users because of ongoing international security concerns around the app, experts have said, as sign-ups appeared to go live once again.

One security expert said the platform could have been managing user sign-ups over the last week “strategically” because of “international security concerns”.

The chatbot sparked upheaval in the US stock market last week after emerging as a possible rival to US AI giant ChatGPT, despite claiming to have been developed for less than ten million US dollars – a fraction of the billions being spent on AI tech by the US giants.

As it upended the markets and the perceived landscape of the AI sector, DeepSeek jumped to the top of app download charts around the world, before suddenly limiting new user registrations to only those with a Chinese phone number.

The company said at the time this was because of “large-scale malicious attacks” on its services and added a warning banner to the log-in page of its website.

This banner has now been removed and it is possible to complete a new sign up and log into the service.

However, a number of security concerns have been raised about the app, most notably because it stores its data in China, raising fears from critics about data being harvested for Chinese intelligence purposes.

ChatGPT maker OpenAI has also accused DeepSeek of using OpenAI’s models to train their own.

Italy has already blocked the app, while Australia has banned it from government devices, with the US now considering similar measures, according to reports.

Bill Conner, chief executive of automation firm Jitterbit, and a former security advisor to the UK and US governments told the PA news agency: “It’s possible that DeepSeek has been managing user sign-ups strategically due to international security concerns raised by the US with its proposed bill to ban the technology as well as Australia’s ban.

“Recent reports have highlighted potential security vulnerabilities in the model, as well as concerns about data transmission practices directly to China Mobile, a Chinese state-owned telecommunications company that has been banned from operating in the United States.

“Additionally, the rapid surge in user demand may be straining their infrastructure and undermining DeepSeek’s claims of the model’s capabilities.

“Given these factors and the broader geopolitical considerations, DeepSeek’s approach to limiting registrations to mainland China numbers seems to have been a strategic decision to navigate regulatory scrutiny and international relations.”

Chris Anley, chief scientist at cyber security firm NCC Group, noted DeepSeek’s shutdown of registrations was the “most logical way to defend its systems”.

“If attackers can’t get a valid account, then they can’t make requests and consume lots of resources, so in turn can’t disrupt services,” he told PA.

However he urged users to exercise caution when using any AI chatbot, as many hackers were using their growing popularity to trick users into sharing personal information.

“When handling chatbots like DeepSeek, always make sure you are using the legitimate app for the chatbot and be careful what information you give it,” he said.

“Third party apps could be malware which harvest your data, so avoid sharing sensitive data like personal or financial credentials.”