Apple and Google receive Deepseek data risk warning from German regulators

A German privacy regulator has cautioned Apple and Google over the availability of the Chinese AI tool Deepseek on their app stores, citing concerns that the platform may be sharing user data with the Chinese government.

Berlin asked Deepseek to withdraw its app from German app stores in May or introduce adequate data safeguards to protect local users’ data.

Europe’s privacy watchdog had also previously raised concerns about the Chinese platform, warning that the chatbot could face further enforcement actions.

Deepseek has been in European countries for months

In early February, privacy regulators across Europe deliberated on Deepseek’s data security risks, shortly after Italy barred the chatbot over its opacity on how it handles user data. France, the Netherlands, Belgium, and Luxembourg also raised alarms over the app’s data collection practices.

The meeting concluded with countries warning that further action could be taken against the platform, and this Friday, Germany put pressure on Apple and Google to remove the app from their stores. This comes after Deepseek did not comply with the German government’s request to remove the platform from app stores.

On Friday, Berlin data protection commissioner Meike Kamp commented, “Chinese authorities have far-reaching rights to access personal data. DeepSeek users don’t have enforceable rights and effective legal remedies available to them in China, like they’re guaranteed in the European Union.”

He added that both Apple and Google must promptly take note of the notice and determine how they will comply.

The German regulator, however, had the option of imposing a fine on Deepseek, instead of warning the two companies. Still, Kamp opted against it, knowing the penalty can’t be enforced in China.

Deepseek has not made it clear when it plans to release its R2 model

Meanwhile, Deepseek has yet to clarify the release of its R2 model, which many are waiting on, especially after the success of its R1 language model, touted by the Chinese startup as a cost-effective rival to larger US systems.

On Thursday, CEO Liang Wenfeng stated that he was unsatisfied with R2’s current performance, which could explain the delayed release. Users had hoped for a May rollout. 

However, staff working in Chinese cloud platforms providing DeepSeek models to businesses revealed that R2’s release was stalled due to difficulties obtaining Nvidia server chips due to tightened US export policies. They believe that increased demand for the new model would also strain cloud firms that need Nvidia chips to run AI models. The platform’s R1 model still heavily depends on Nvidia’s H20 chips. 

So far, Deepseek has maintained contact with Chinese cloud companies, providing direction on technical requirements to help them prepare for model deployment and distribution. 

Nevertheless, China’s Cloud companies are not the only ones in for a loss. Speaking on the export restrictions, Nvidia’s billionaire chief executive, Jensen Huang, claimed they may have to take a $5.5 billion loss on the H20 AI chips meant for the Chinese market to maintain compliance.

He asked the US government to reconsider its chip ban and allow the company to win back the Chinese market. He argued that if the US wants to lead in tech, they have work to maximize its output, not lower it, hinting that other countries could easily overtake it.

Currently, Nvidia controls only about 50% of the Chinese market, down from about 95% at the start of Biden’s administration in 2021.

Your crypto news deserves attention - KEY Difference Wire puts you on 250+ top sites