OpenAI Says China Could 'Compel' DeepSeek to Cause Harm, Calls for Ban

INA-  SOURCES
ChatGPT maker OpenAI has hit out at Chinese AI lab DeepSeek, alleging it “could be compelled by the Chinese Communist Party (CCP) to manipulate its models to cause harm."

In a letter to the US Office of Science and Technology Policy, OpenAI calls for a ban on China-produced equipment and AI models "that violate user privacy and create security risks such as the risk of intellectual property (IP) theft" in "Tier 1" countries like the UK, Canada, and Germany.

The firm alleges that “there is significant risk in building on top of DeepSeek models in critical infrastructure and other high-risk use cases” due to the potential for Chinese government interference. OpenAI compares the risk to that posed by Chinese telecom and consumer tech giant Huawei, which is heavily sanctioned in the US and much of Western Europe.

In trying to ban TikTok, the US government also argued that parent company ByteDance could be compelled to hand over control to the CCP.

DeepSeek founder Liang Wenfeng reportedly met with Chinese leader Xi Jinping during an AI summit. But, to be fair, OpenAI CEO Sam Altman also appeared at the White House alongside President Trump to champion his company's participation in Project Stargate.

Banning DeepSeek and other Chinese AIs would likely benefit OpenAI's business, so the letter may be a tad self-serving. OpenAI voicing concern about the "risk of intellectual property (IP) theft" is also a little rich given that it's currently embroiled in several lawsuits that accuse it of scraping copyrighted content without permission to train its own models.

DeepSeek hit the headlines at the start of this year, drawing attention for its impressive performance and low cost, with some commentators dubbing it a “Sputnik moment” in the technology race between China and the US. It quickly became one of the most downloaded AI apps in the US.

OpenAI's letter was in response to the AI Action Plan, a US government initiative aimed at maintaining American dominance in AI, which called for proposals from AI firms like OpenAI.

OpenAI also petitioned for several other measures it claims will speed up American AI dominance. These include making it easier for government agencies to use the services of "frontier AI" companies and applying "fair use doctrine" to AI training data.

"While America maintains a lead on AI today, DeepSeek shows that our lead is not wide and is narrowing," says Chris Lehane, OpenAI's VP of global affairs.

Regardless of OpenAI's requests, many parts of the US government are already pushing for a hardline approach on DeepSeek. The House introduced a bill last month that would ban DeepSeek on government devices, with Rep. Josh Gottheimer (D-N.J.) calling the AI model "a five-alarm national security fire." Meanwhile, DeepSeek has already been banned on government devices in Texas and New York.

Last week, the Wall Street Journal reported that the current administration is mulling an outright ban on DeepSeek in the US, though discussions are reportedly "still at an early stage" and a crackdown may only cover government devices. Other countries have already blocked DeepSeek on government devices, including South Korea, Taiwan, and Australia.



SOURCE: PC MAG