OpenAI claims that competitors, including Chinese firms, are using its technology to rapidly develop new AI systems. The emergence of DeepSeek, a Chinese AI app, highlights this concern. DeepSeek reportedly rivals ChatGPT at a much lower cost, raising questions about unauthorized usage.
DeepSeek’s Rise Sparks Security Concerns
Bloomberg reported Microsoft is investigating potential unauthorized use of OpenAI’s data by DeepSeek. OpenAI’s concerns align with remarks from David Sacks, the White House “AI and crypto czar,” who warned of a practice called knowledge distillation. He accused DeepSeek of extracting valuable insights from OpenAI’s models to enhance their own system.
“DeepSeek has distilled knowledge from OpenAI models,” Sacks stated on Fox News. “Major US AI companies will take measures to prevent this, slowing copycat models.”
The US government has already enforced measures to safeguard its AI innovations. Export restrictions block China’s access to advanced chips and prioritize investments within the US. Howard Lutnick, Trump’s nominee for Commerce Secretary, echoed these concerns during his confirmation hearing. He suggested more steps are needed to shield US firms from unauthorized replication.
“Our current export controls resemble a whack-a-mole system,” Lutnick said.
OpenAI emphasized the need for close collaboration with the US government to protect leading AI technologies from misuse. “Chinese and other firms are constantly trying to distill our models,” the company stated.
Distillation and Deceptive Cost Claims
Naomi Haefner, a technology management professor at the University of St. Gallen, questioned DeepSeek’s claim of training AI models inexpensively. “It’s unclear if DeepSeek built its models independently,” Haefner said. OpenAI suspects DeepSeek misappropriated large amounts of data to lower costs.
“If true, claims about training the models cheaply are deceptive,” Haefner added. “We’ll need someone to replicate the approach to confirm this cost-efficiency.”
Crystal van Oosterom, AI Venture Partner at OpenOcean, noted DeepSeek built on publicly available research. However, she acknowledged that the ethical boundaries around “building on” others’ work remain unclear, particularly in AI. US AI firms have also faced accusations of violating intellectual property rights.
National Security and Ethical Concerns
The White House is assessing the national security implications of DeepSeek’s development. White House press secretary Karoline Leavitt confirmed that the National Security Council is investigating possible risks. Leavitt referenced President Donald Trump’s warning that DeepSeek should act as a wake-up call for the US tech sector.
The US navy responded by banning its members from using DeepSeek apps due to potential security and ethical risks. According to CNBC, the navy sent an internal email warning staff against the app’s data collection practices and its ties to servers in China.
Cybersecurity experts have advised users to avoid DeepSeek due to potential risks tied to data storage and usage. Meanwhile, DeepSeek claims to be under cyberattack, citing large-scale disruptions that have forced temporary restrictions on new user registrations. A message on DeepSeek’s website attributes the delays to ongoing attacks.
