Ed Husic, Australia‘s science minister, has made headlines as the first official from a Western government to express privacy concerns regarding DeepSeek, the Chinese chatbot currently stirring significant unrest in the markets and the technology sector.
Chinese technology companies, including Huawei and TikTok, have faced ongoing allegations regarding their connections to the Chinese government. Concerns persist that these ties may result in the potential harvesting of personal data for intelligence purposes.
Donald Trump described DeepSeek as a “wake-up call” for the United States, though he did not indicate that it posed a threat to national security. Instead, he suggested that it could be beneficial if it reduced costs.
On Tuesday, Husic expressed to ABC News that numerous questions still linger, particularly regarding “data and privacy management.”
“He emphasized the importance of caution, stating that such issues require careful consideration,” he added.
DeepSeek has yet to reply to the BBC’s inquiry for comment; however, UK and US users appear to be engaging without hesitation.
DeepSeek has surged to the forefront of app stores in both nations, with market analysts at Sensor Tower reporting that it has achieved 3 million downloads since its launch.
In a remarkable surge, up to 80% of these downloads have occurred within the last week, indicating a pace three times faster than competitors like Perplexity.
DeepSeek gathers various types of data, though specifics on the categories and nature of this information remain undisclosed.
DeepSeek’s privacy policy reveals that the platform gathers extensive personal information from its users and stores it “in secure servers” in China.
This could encompass:
The email address, phone number, and date of birth provided during the account creation process
All user contributions, encompassing both text and audio formats, along with chat records
The term “technical information” encompasses a variety of data points, including your phone’s model and operating system, your IP address, and even your keystroke patterns
The statement indicates that this information enhances DeepSeek, focusing on improving its safety, security, and stability.
The information will subsequently be disseminated to various parties, including service providers, advertising partners, and the corporate group, with retention lasting “for as long as necessary.”
“Concerns are mounting regarding the technological capabilities of DeepSeek, particularly about its privacy policy,” stated Lauren Hendry Parsons, digital privacy advocate for ExpressVPN.
She emphasized a particular section of the policy stating that data may be utilized “to help match you and your actions outside of the service,” which she argued “should immediately raise concerns for anyone worried about their privacy.”
Despite the app’s extensive data collection, experts highlight that its privacy policies closely resemble those that users may have previously accepted for competing services such as ChatGPT and Gemini, as well as various social media platforms.
“Any AI model that is openly accessible, whether through a web or app interface, such as DeepSeek, allows the prompts or questions posed to the AI to be collected by the developers of that model, along with the corresponding answers,” stated Emily Taylor, chief executive of Oxford Information Labs.
“Individuals engaged in confidential or national security sectors must remain vigilant regarding these risks,” she stated in an interview with the BBC.
Dr. Richard Whittle of the University of Salford expressed his apprehensions regarding the app’s data and privacy while also noting that the models employed in the United States have numerous issues.
“Consumers must exercise caution, particularly amid the excitement and anxiety surrounding the latest trending app,” he stated.
The Information Commissioner’s Office, the UK’s data regulator, has called on the public to remain vigilant regarding their rights concerning using their information in training AI models.
In response to inquiries from BBC News regarding the Australian government’s concerns, the organization issued a statement emphasizing the importance of generative AI developers and deployers providing individuals with clear, concise, and easily accessible information about the utilization of their personal data.
Additionally, it highlighted the necessity for effective processes that empower individuals to exercise their information rights. We will persist in collaborating with stakeholders to advance effective transparency measures, and we will not hesitate to take action when our regulatory expectations are overlooked.
