The Army’s technology acquisition office announced today two new initiatives as part of its 500-day artificial intelligence implementation plan, focusing on both the testing of cutting-edge technology for soldiers and the defense against potential enemy AI usage.
Young Bang, the principal deputy to the assistant secretary of the Army for Acquisition, Logistics & Technology (ASA(ALT)), introduced the initiatives using new hashtags: #BreakAI and #CounterAI.
Testing and Validation with #BreakAI
#BreakAI is designed to test algorithms through traditional government „testing and evaluation” and „verification and validation” processes, ensuring that AI systems are fully operational and error-free when deployed to soldiers in the field. Bang emphasized the challenges of testing AI, especially as the Army moves toward artificial general intelligence (AGI), questioning how one can effectively test systems when the outcomes and behaviors are unpredictable.
“A lot of potential exists, but experts have warned about the difficulty in understanding how AI arrives at its answers, comparing it to a ‘black box,’” Bang noted. He highlighted the need for industry collaboration in implementing #BreakAI, specifically in developing tools for testing AI capabilities and soliciting feedback from industry partners.
“#BreakAI is about ensuring that our AI models perform accurately so that when they are in the hands of soldiers, we can be fully confident in their reliability and effectiveness,” stated Jennifer Swanson, deputy assistant secretary of the Army for data, engineering, and software.
Defending Against AI Threats with #CounterAI
In contrast, the #CounterAI initiative focuses on defense strategies. Its goal is to secure Army platforms, algorithms, and capabilities from potential attacks and threats. Swanson pointed out the urgency of this initiative, noting the substantial investments made by other countries in counter-AI technologies that pose a significant risk to the U.S.
„We know that we’re not the only ones investing in this area; many countries that are major adversaries to the United States are actively pursuing countermeasures,” she added.
Integrating Initiatives for Comprehensive Coverage
The #BreakAI and #CounterAI initiatives complement the previously announced #DefendAI initiative, introduced in June as part of the Army’s 500-day AI strategy. The #DefendAI initiative aims to create a layered AI defense framework to mitigate risks associated with third-party algorithms while working to operationalize industry AI. While #DefendAI focuses on addressing issues that arise from using outsourced AI, #CounterAI is more concerned with adversarial threats.
“I think there will be potential linkages and lessons learned from #DefendAI that we can apply to #CounterAI,” Swanson noted.
Despite seeking industry collaboration for these initiatives, the Army is cautious about sharing its findings publicly, as rival militaries are keenly pursuing their own AI advancements.
“We won’t disclose everything publicly, as operational considerations must be taken into account. There will be things we share in forums like this, but there are also matters we must keep confidential to prevent adversaries from understanding our countermeasures,” Swanson explained.
