Trump Administration Wants To Ban DeepSeek From The US

Summary: The Trump administration is thinking about banning China’s DeepSeek AI from being used in the US. It is also trying to hamper the company’s efforts by restricting it from buying NVIDIA’s AI chips.

According to a report from The New York Times, the Trump administration is considering placing a ban on DeepSeek in the US. It also seeks to hinder the company’s AI development efforts by limiting its buying of NVIDIA’s AI chips.

DeepSeek has exploded in popularity over the past few months. The company burst onto the scene with its own AI model that boasted capabilities similar to those of other mainstream models, such as OpenAI’s ChatGPT. It even claimed to have trained it using just $6 million, a fraction of what other AI companies are spending. This resulted in the AI market temporarily crashing as many started to wonder if all this spending was truly necessary.

A potential ban

However, it looks like DeepSeek’s stay in the US and its future as a whole could be in jeopardy. On Tuesday, the White House made another attempt at restricting NVIDIA’s chip sales to China. This would further tighten rules originally put in place by former US President Joe Biden. Given that NVIDIA supplies the hardware that powers the majority of AI models out there, this is a huge deal.

According to the report, sources say the Trump administration is discussing penalties that would block DeepSeek from accessing US technology. There’s also talk of placing an outright ban on American users from using DeepSeek in the US. However, whether this is out of legitimate concerns about DeepSeek’s security or just another way to force China to agree to its demands, is unclear.

DeepSeek’s deep problems

DeepSeek impressed many when it first arrived, especially with its claim that it trained its model on just $6 million. However, security researchers quickly realized that the AI model had all sorts of flaws and vulnerabilities. Attackers have tricked DeepSeek into doing and revealing things it shouldn’t, like generating malware in response to user prompts.

Later reports revealed that DeepSeek’s parent company spent far more than the $6 million it initially claimed. While that figure was technically accurate, it only covered a portion of the overall costs. The company also faces accusations of using data from other AI models without permission to train its own systems.

We realize the irony of the last statement. Many companies developing AI have been accused of using copyrighted material to train their models, so the hypocrisy is not lost on us. That being said, the US has not made any final decisions yet on banning DeepSeek. But if you are living in the US and rely on DeepSeek heavily, it’s time to look for a new model.