The Political Imperative for Regulating AI in National Defense

• Artificial intelligence (AI) enhances military operational speed but introduces uncontrollable risks regarding malfunction and ethical judgment.
• Automating military decisions does not improve their quality; instead, it relies on statistical probability rather than objective certainty, leading to dangerous errors.
• The "human-in-the-loop" concept is increasingly compromised by operational machine speeds and "automation bias," where human oversight becomes a formality.

Current Risks and Systemic Challenges

• Modern AI operates via pattern recognition, making its internal logic inherently opaque, even to its creators.
• Recent incidents involving autonomous agents infiltrating civilian and secure systems highlight the danger of these technologies being integrated into weapons systems.
• Traditional arms control, based on counting tangible units like tanks or planes, is ineffective against AI and cyber-based capabilities that are invisible, easily proliferated, and hard to verify.

Strategic Recommendations

• Major military powers must establish robust communication channels and crisis hotlines to prevent escalations caused by AI-driven miscalculations.
• International frameworks must move beyond voluntary commitments to implement binding bans on fully autonomous lethal weapons systems.
• Europe should take the initiative in creating global standards, positioning itself alongside neutral mediators to offer a credible alternative to the rivalry between Washington and Beijing.
• Policy must prioritize human wisdom over technological acceleration, explicitly defining domains where machines must never be permitted to make life-or-death decisions.