A group of Republicans led by Senator Ted Budd is calling for an inquiry into DeepSeek, a Chinese company, due to potential national security threats posed by its open-source AI model R1.
Contents of the Letter to the U.S. Commerce Department
On August 1, a group of senators sent a letter to the U.S. Commerce Department expressing concerns that DeepSeek's AI model may leak U.S. user data to assist China's military and surveillance efforts. The letter was made public on August 5.
Safety Issues and Potential Threats
The senators highlighted that the R1 model offers open-source access, allowing programmers to modify its code. They noted a high potential for generating harmful content. Research from Wiz Research found a publicly accessible database belonging to DeepSeek containing sensitive information, including chat histories and secret keys. Instances of dangerous content generated by the AI included instructions for self-harm campaigns among teenagers and guidance for developing bioweapons.
China's Responses and Next Steps for the U.S.
The senators also expressed concern that China has found ways to embed open-source AI models into applications despite limited access to advanced computing resources. China's President Xi Jinping emphasized the need for a strong application-oriented approach in the country's AI sector. In response to U.S. concerns, the Chinese Foreign Ministry assured that it protects user data and has never required companies to collect information unlawfully.
The situation with DeepSeek highlights the importance of oversight over open-source AI models and the need to protect user data from potentially harmful influences.
