United States Invests in ‘Agentic AI’ Advanced Model Development

The United States has signed contracts with advanced artificial intelligence developers to create a new model of “agentic AI” that will be able to act independently to achieve its goals.

This was reported by the Pentagon’s Chief Digital and Artificial Intelligence Officer (CDAO).

The defense body, responsible for researching artificial intelligence models, has announced three equivalent contracts worth $600 million with leading developers Anthropic, Google, and xAI.

The companies’ developments will form the basis of the army’s Enterprise LLM Workspace platform, which will allow the secure use of several commercial AI models in one working environment. Later, this tool will be implemented in a wide range of offices of the Department of Defense.

The CDAO wants to go beyond the known generative AI models that are capable of generating a conditional action plan by creating the next level of development of agentic AI technology, which will be able to independently understand the goal, build an action plan to achieve it, and execute it, adapting to changes in conditions.

“The grants … will allow the Ministry to utilize the technology and talents of American advanced AI companies to develop ‘agentic AI’ workflows in various fields of activity,” the CDAO press release says.

The creation of such a system will be a breakthrough that will allow the new development to assign tasks that previously required human participation.

In particular, not only management and control tasks, but also elements of combat missions may be partially or fully automated, the press release said.

“The use of commercially available solutions … will accelerate the use of advanced artificial intelligence in the field of combat operations, as well as in intelligence, business and corporate information systems,” the department said.

The investments of the DoD make up only a small share of the funding for the development of the above-mentioned companies. For example, OpenAI alone reported annual revenue of $10 billion last month, and in March it raised a record $40 billion from third-party investors.

However, the allocation of more than half a billion dollars for the development of advanced AI technologies is a clear sign of the Pentagon’s interest in their implementation in the US defense system.

The contract was preceded by Google’s public renunciation of its previously adopted ethical principle of not using its own AI technologies for weapons and surveillance.