Google and Microsoft: Trump Tests AI in Secret

Trump and the Tech Giants: Anticipatory Oversight

Donald Trump has decided to take a proactive stance in response to the rise of the most powerful artificial intelligences on the market. Google, Microsoft, and xAI have agreed to submit their future models to government evaluations before public deployment. This initiative, stemming from the former president, aims to enhance national security and prevent risks related to cybersecurity, misinformation, and military uses.

Traditionally, new AI models would appear on the Internet almost instantly. Now, they may first pass through the offices of the U.S. government. Under Trump's influence, Washington seeks to test certain AIs before their public launch. Google DeepMind, Microsoft, and xAI are already engaged in this process, which resembles both a precautionary measure and a demonstration of technological power.

The Role of CAISI in AI Evaluation

The Center for AI Standards and Innovation (CAISI), an agency of the Department of Commerce, is tasked with evaluating the capabilities of so-called frontier AIs. Until now, the center has already been collaborating with companies like OpenAI and Anthropic. Now, Google DeepMind, Microsoft, and xAI are officially joining this program, allowing the U.S. government to test the most advanced AIs before they hit the market.

The U.S. government claims it wants to better understand the risks associated with these technologies. Chris Fall, director of CAISI, emphasizes the need for an independent and rigorous evaluation of the most advanced models. The arguments for this evaluation include national security, cybersecurity, misinformation, and military uses.

Towards Enhanced AI Regulation

According to a report from the New York Times, Donald Trump is considering signing a presidential decree to further strengthen oversight of AI technologies. The goal is to bring together tech leaders and government officials to more directly oversee the development of new AIs.

This collaboration between Trump and xAI, Elon Musk's company, is particularly noteworthy. Musk, who regularly criticizes the excesses of "woke" AI and advocates for more open models, is now agreeing to collaborate with U.S. authorities. This alliance underscores the growing geopolitical importance of AI.

AI: A Major Geopolitical Issue

As global competition intensifies, the United States seeks to maintain its lead over China. Tech companies, for their part, are trying to reassure governments while pursuing their business ambitions. The question remains who will truly control the AIs of tomorrow: the companies that create them, the governments that evaluate them, or a few billionaires convinced they are building the future of humanity from their ultra-secure open spaces?