US Government Website Used an AI Search Tool Made in China

A United States government website used an AI search tool made by a Chinese company called Alibaba. The website, called the Federal Register, lets people read and comment on new government rules. The AI tool, known as Qwen, was available on September 16 but was taken down the same day after people noticed it on social media. This happened even though the FBI had accused Alibaba of illegally copying technology from an American AI company.

The Federal Register is run by the National Archives, a federal agency that keeps important government records. The site offers a daily list of newly proposed regulations, which are official rules the government is thinking about making. The Qwen tool was offered as a way for users to search through public comments on those rules. It is not clear exactly when the tool was first added to the website.

The FBI recently accused Alibaba of 'malicious' copying of technology from Anthropic, a well-known American AI company. US officials said Chinese AI companies were stealing American technology on an 'industrial scale.' The Chinese embassy in Washington called those accusations 'unfounded,' meaning they said there was no proof. The FBI chose not to comment further after the Federal Register story came out.

Qwen is what experts call an 'open-weight' model. This means that the key parts of the system are made available to the public. Developers can download it, change it, and use it for their own purposes, often for less money than using AI tools from American companies like Anthropic or OpenAI. Because of this, many organizations around the world use it.

Experts said the use of Qwen on the Federal Register website likely did not create a big security risk. That is because the website only deals with information that is already available to the public. Georgetown Law Professor Anupam Chander said there did not appear to be an immediate cybersecurity danger, though he noted it depends on how the model was trained. Senator Mark Warner of Virginia said the real risk would depend on whether US data was sent to systems controlled by Alibaba.

Still, many officials were upset about the situation. US Representative John Moolenaar, who leads the House China Committee, said no federal government group should ever use a Chinese AI model. He warned that doing so could make the US more dependent on China for important technology. Daniel Castro, the president of a technology research group, called it one of the most surprising things he had ever seen.

The story comes at a time when the US and China are already in a tense moment. Leaders from both countries were set to meet soon, and the FBI's accusations about Alibaba had already stirred up debate. The US is also having a big discussion about whether the government should create new safety rules for AI or whether doing so might slow the country down in the global AI race. The Qwen situation added a new twist to that ongoing debate.

This is not the first time Chinese AI tools have raised concerns in the US. Earlier this year, committees in the House of Representatives sent a letter to Airbnb, the home-rental company, because it also uses Qwen. Lawmakers wanted to know more about how Airbnb uses the tool and whether it creates risks for Americans. The investigation is still ongoing as the government tries to figure out how to handle the growing use of Chinese AI technology inside the United States.