Congress Might Ban DeepSeek

Well, here we go again. With TikTok’s future in the United States still uncertain, parts of Congress are already moving toward banning DeepSeek, a Chinese ChatGPT competitor that quickly usurped ChatGPT on the App Store last week.

A new bill introduced by Republican Senator Josh Hawley of Missouri would “prohibit United States persons from advancing artificial intelligence capabilities within the People’s Republic of China.” In other words, it would effectively ban DeepSeek, despite not mentioning the AI chatbot by name.

Under the bill, anyone convicted of “the import from or export to China of artificial intelligence technology” could be sentenced with up to 20 years in prison, plus an up-to-$1 million fine for individuals and an up-to-$10 million fine for businesses.

Are you likely to be punished for using DeepSeek as a private citizen? Well, the bill does leave room open for it, so practice caution if it does become law. More realistically, much like with piracy laws, this is probably aimed at distributors and developers, so if it goes through, expect to see DeepSeek and other Chinese AI leave the App Store.

The bill is also likely so broadly worded as a result of DeepSeek’s open-source nature, as it’s currently easy for developers to simply download its models and incorporate them into their apps. The legislation would seek to curb such a practice, although it would also put future international collaboration, even the reading of Chinese research papers about AI, in jeopardy.

In a statement, Senator Hawley said “Every dollar and gig of data that flows into Chinese AI are dollars and data that will ultimately be used against the United States.”

Senator Hawley isn’t the first to express concern about DeepSeek. The AI has already been banned for use in the Navy due to “potential security and ethical concerns,” as well as on Texan government devices. OpenAI also toldTthe Financial Times that it has evidence that DeepSeek might have been built on the back of OpenAI models, which while a common practice in AI development, goes against OpenAI’s terms of service once the resulting models are distributed. There’s also already been a high profile data leak, and DeepSeek’s terms of use involve agreeing to having your keystrokes collected.

At the same time, the strict language of the law does have experts worried about the future of AI development and oversight, as it could kneecap collaboration and AI transparency. The Electronic Frontier Foundation’s Director of AI and Access-to-Knowledge Legal Projects, Kit Walsh, told 404 Media that “The bill threatens the development and publishing of AI advancements in the United States…the government has argued that merely publishing information on the internet counts as an export, and interpreting this law in such a way would further solidify the dominance of proprietary AI over open or academic research.” In other words, American tech companies would be incentivized to be even more insular about how their AI works, which could affect far more than Chinese competitors.

View the full article