The US Central Command has also attempted to use an AI recommendation engine to suggest optimal weapons for operations and create attack plans, but this technology has often fallen short. Every AI-involved step is checked by a human. Amid fears of falling behind technologically advanced adversaries, the Department of Defense is increasing efforts to integrate and test AI's warfighting capabilities. The agency's chief digital and artificial intelligence officer, Craig Martell, emphasized the need to responsibly adopt generative AI models while mitigating national security risks.
Key takeaways:
- The US Department of Defense has used machine learning algorithms to identify targets in over 85 air strikes in Iraq and Syria this year, continuing a practice started with Project Maven in 2017.
- US Central Command has utilized these algorithms in real campaigns, particularly after Hamas' surprise attack on Israel in the previous year.
- The AI technology has been used to identify enemy rockets, missiles, drones, and militia facilities, but attempts to use it for weapon recommendations and attack planning have frequently fallen short.
- The DoD is increasing efforts to integrate and test AI's warfighting capabilities, driven by concerns of falling behind more capable adversaries and the potential national security risks posed by poorly managed training data.