As U.S. companies reel from sticker shock in the pricing of token-based artificial intelligence, some are turning to cheaper Chinese models, sparking alarm among U.S. lawmakers and national security officials.
CNBC reported on Wednesday (July 8) that the Chinese-developed model is attracting attention among American companies because it narrows the performance gap with the American model and is cheaper to use.
As of Tuesday (July 7), the share of tokens sent to Chinese models from US companies using OpenRouter reached 45%, up from 11.5% at the beginning of the year, according to a report from One Stop API for AI Models.
The trend is raising concerns among U.S. officials about the national security implications, CNBC reports.
“The growing use of Chinese AI models by U.S. companies raises serious concerns,” a State Department spokesperson said, according to the report. These “AI models are designed to advance the Chinese government’s cause and censor and reflect dissent.” [Chinese Communist Party] ideology and values. ”
President Donald Trump in April accused Chinese companies of waging an “industrial-scale campaign” to steal secrets from U.S. AI systems and said he would consider ways to hold foreigners accountable, according to reports.
Congress is also considering curbing the use of Chinese models by U.S. companies. The House Homeland Security Committee and the Select Committee on China announced in April that they would jointly investigate the expanded adoption of Chinese-developed AI models, according to the report. As a first step, the chairs of these committees sent letters to Cursor and Airbnb warning them about the “use or exposure” of risks from AI developed in China.
“The Chinese Communist Party is no longer just trailing behind us in artificial intelligence, it is racing to close the gap in some of the precise capabilities that will shape the future of cybersecurity,” said New York-based Chairman Andrew Garbarino of the U.S. House of Representatives Homeland Security Committee, according to the report. “Recent reports that China’s indiscriminate heavyweight models may rival U.S. leading models in certain vulnerability-finding and cybersecurity tasks are deeply concerning.”
According to the report, Airbnb “overwhelmingly runs its AI activities on a model originating from the United States.” The company added that it “uses a limited number of Chinese-origin models, all of which are open source and run only through approved US-based service providers.”
Ironically, the Chinese government shares U.S. officials’ concerns, albeit for national security reasons. Reuters reported on Tuesday that Chinese authorities have held meetings with domestic tech companies over the past month about potentially restricting foreign access to China’s cutting-edge AI models, including some that have not yet been released.
Officials discussed making the leak or theft of proprietary AI technology a crime under China’s strict national security laws, the report said. It also raised the possibility of new measures restricting who can support domestic AI startups.
The increased use of the Chinese model since the beginning of this year coincides with major US AI companies switching from a subscription payment model to a pricing model based on the number of tokens in a prompt and the corresponding number of tokens generated. As a result, costs have skyrocketed for companies that have encouraged employees to take full advantage of AI, setting off a scramble to find lower-cost models.
In addition to investigating ways to curb the use of Chinese models by American companies, both House committees are also looking into whether the United States is taking sufficient steps to counter the proliferation of Chinese models, CNBC reports.
“The committee is also examining whether the United States has a sufficient open-weight AI strategy to ensure that U.S. companies and cyber defenders are not forced to choose between expensive or limited U.S.-made models and cheaper, more capable U.S.-made models.” [Chinese]”We developed an alternative,” one committee aide said, according to the report.
Tennessee’s Andy Ogles, chairman of the Cybersecurity and Infrastructure Protection Subcommittee, called for a “serious strategy” to ensure the U.S. model is a “true replacement” for the Chinese model, the report said.
“If China has a cheaper, more capable and easier option for AI models, the rest of the world will build on it,” he said, according to the report.
For all of our coverage of PYMNTS AI, subscribe to our daily subscription AI Newsletter.
