With AI hype outrunning reality, DoD AI chief Craig Martell told lawmakers his office is "building what we're calling a maturity model" to assess what generative AI really can and cannot do.
Because it needs much less electricity per computation than current chips, the new hardware could take AI out of big data centers and onto drones, robots, and other small platforms at the “tactical edge.”
The US government hopes this will be the first of many annual meetings of the countries that signed on to the US “Political Declaration” on military AI last year, sharing model policies and best practices on everything from combat robots to back-office algorithms.
Plumb, the current deputy under secretary of defense for acquisition and sustainment, has held high-powered jobs from Google to the National Security Council.
“I love AI. I want lots of AI,” Dave McKeown told Breaking Defense. But, so far, neither government nor industry has developed artificial intelligence that can really help with cybersecurity.
“On an annual basis, we have a CR, half the year’s gone,” the undersecretary for Research & Engineering said. “Our adversary doesn’t have the same constraints."
Recently confirmed Gen. Jim Slife pledged to work on sharing data not only inside the Air Force but with his counterparts at the other armed services.
2023’s Global Information Dominance Experiments were a race to field a “minimum viable” version of CJADC2. The 2024 GIDEs will open up to more input from the services, the Five Eyes allies, and a wide range of contractors.
With official outlets like RT and Sputnik kicked out of many Western countries, Moscow now emphasizes social media — exploiting TikTok, X, and the explosion in generative AI, according to a new report.
“The Minimum Viable Capability for CJADC2 [Combined Joint All Domain Command & Control] is real and ready now,” said Deputy Secretary Kathleen Hicks. “It's low latency and extremely reliable.” But budget gridlock blocks deployment.
The US military can quickly kludge together custom solutions for urgent problems, said Pentagon AI chief Craig Martell. But it needs a more systematic approach to turn those quick hacks into stable, enduring AIs that can actually talk to each other.
Last fall, both Russia and Ukraine boasted of fielding drones that tracked targets using AI algorithms. Now new expert analysis suggests that neither side got it to work well enough for war — but the US and China might.
With Chinese wargames growing ever larger and more realistic, Adm. Sam Paparo said traditional intelligence “indications and warning” can’t tell if they’re just practicing or actually about to invade Taiwan. His solution: surveillance drones feeding AI analysis.
It’s tempting to use AI to micromanage military operations from an all-seeing HQ. But experts warn the chaos of war means AI is best used to empower decentralized initiative — an area where the democratic West just might have the edge.
Budget dysfunction in DC can disrupt Pentagon programs and leave allies hanging. The US needs to reform its budgeting process to support its leadership role, two RAND scholars argue.