The US government approves contracts with OpenAI, Google, and Anthropic for the use of AI in the public sector.

247 - The United States government has officially authorized the use of artificial intelligence tools developed by OpenAI, Google, and Anthropic in civilian federal agencies. This information was originally reported by Bloomberg.

The measure was announced this Tuesday (5) by the General Services Administration (GSA), which included the ChatGPT (from OpenAI), Gemini (from Google) and Claude (from Anthropic) models in the Multiple Award Schedule, the country's main government procurement system. With this, civil public bodies will be able to immediately access these AI systems, without having to sign individual contracts with each supplier.

The approval of the contracts comes days after President Donald Trump signed three executive orders to regulate the use of artificial intelligence in public service. Among the guidelines is the requirement that only systems "free from ideological bias" be adopted. According to the GSA, compliance with the order will be the responsibility of each agency.

“But at the same time, this is a race, right? And as the president said, we are going to win this race,” said Josh Gruenbaum, commissioner of the Federal Procurement Service, referring to the global competition for technological dominance.

Tools evaluated by technical criteria

According to the GSA, the AI ​​models underwent rigorous performance and security testing before being authorized. The agency did not detail the contractual terms, but historically it has negotiated large discounts with vendors such as Adobe, Salesforce, and Google itself.

The inclusion of the three companies does not conclude the process. Other suppliers may also be qualified in the future. “We are not choosing winners or losers here. We want as many tools as possible available to all federal public servants, so that they can be as productive as possible,” explained GSA Deputy Administrator Stephen Ehikian. “Different tools will be suitable for different types of use.”

The inclusion of these technologies on the federal platform represents a transition from experimental use—restricted to pilot projects and the defense sector—to widespread adoption in the civilian public sector. The Department of Defense had already signed contracts with OpenAI and xAI, Elon Musk's company, in agreements separate from those announced this week.

Applications and limitations

Among the agencies interested in adopting the models are the Treasury Department and the Office of Personnel Management (OPM). According to Scott Kupor, director of the OPM, the models could be used to develop customer service chatbots and to summarize thousands of contributions received in public consultations on regulations—tasks that previously compromised regulatory review deadlines.

Kupor cautioned, however, that implementing these technologies will require skilled hires: "We are probably missing out on hiring people who are highly fluent in modern topics related to AI," he said.

He also highlighted the need for careful planning: "Of course, we can't just throw things at the wall and see what sticks," he stated.