The Pentagon should start a multi-billion dollar artificial intelligence (AI) program with a contractor not among DoD’s usual stable, according to a top GOP lawmaker.

“I’m concerned that there’s an argument right now being made by some in the [Biden] administration that our interests with China overlap when it comes to regulating AI,” Rep. Mike Gallagher (R-Wis.), the chairman of the House Select Committee on the Chinese Communist Party, said at an AI panel at the Ronald Reagan National Defense Forum on Dec. 2 in Simi Valley, Calif.

Gallagher also chairs the House Armed Services Committee’s (HASC) Cyber, Information Technologies, and Innovation (CITI) panel.

“What we need to do is recognize that this isn’t 1975 where we can embark on a set of these [arms control] agreements,” Gallagher said. “It’s more akin to 1945. We need to build these weapons and systems rapidly with our allies. In simplest terms, what can the Pentagon do? Pick winners and losers in this space. Put their thumb on the scale. Have a multi-billion dollar ACAT I program in procurement involving a non-traditional company leveraging AI and autonomous systems. I don’t think we have that right now.  That’s where we in Congress need to provide the flexible funding in the form of multi-year appropriation. We need to encourage the Pentagon to take intelligent risk, but we cannot simply admire the ethical and policy dimensions of this problem at the expense of fielding capability right now.”

During a July hearing of the HASC CITI panel, Scale AI CEO Alexandr Wang said that DoD has not had a pathfinder AI project start since Maven and that it should create new AI pathfinders to drive Pentagon AI innovation (Defense Daily, July 18).

At the Intelligence and National Security Association’s spring symposium in March, Phillip Chudoba, the National Geospatial-Intelligence Agency’s (NGA) associate director of capabilities, said that Maven is the “only performant computer vision, AI/ML capability in the DoD.”

U.S. Air Force Secretary Frank Kendall agreed with Gallagher that generative AI is vital in areas, such as pattern recognition and process efficiencies, but said that knitting together programs that use AI is the most important part of DoD adoption of AI.

Generative AI, able to create new content–text, imagery, audio, and synthetic data, “is being embedded into all of our products and is being used where it provides a competitive advantage,” Kendall said at the Reagan forum. “The government, as much as anything else, needs to not get in the way of that and encourage that. I don’t think we frankly need an AI program, per se. What we do need is find ways to evaluate that technology, become confident in it/have the ability to trust it, and get it into fielded capabilities as quickly as we can.”

Brian Schimpf, the co-founder and CEO of Anduril, told the AI panel at the Reagan forum that letting military services begin experimenting with fielded, AI-enabled platforms is key and that the Air Force’s Collaborative Combat Aircraft (CCA) program is “the brightest spot” at the Pentagon for developing, fielding, and scaling up “next generation” AI technologies.

DoD requested $1.8 billion for AI in fiscal 2024 and, of small business awards last year, one-third were for AI, said Undersecretary of Defense for Research and Engineering Heidi Shyu. Some of the biggest benefits of AI may lie in maintenance cost savings in which AI is able to predict when parts will fail and have spare parts available, she said.

The Air Force AI predictive maintenance effort has included a contract for the B-1 bomber with Silicon Valley’s C3.ai, Inc. [AI]. In April, the company said that the Air Force chose the Rapid Sustainment Office’s Predictive Analytics and Decision Assistant-developed with C3.ai, Inc., as the service’s system of record for Condition Based Maintenance Plus and predictive maintenance

While the DoD fiscal 2024 budget request represents a high point dollar-wise for AI, the Pentagon is spending three times less than China on AI, which the People’s Liberation Army is positioning for use in autonomous drone swarms, adaptive radar systems, and autonomous vehicles, Wang told the HASC CITI panel hearing in July.