Artificial Intelligence in Autonomous Weapon Systems

Abstract

Cyber space, the fifth warfare domain, is benefitting from technological advancements and already accommodating Artificial Intelligence (AI) as its essential part. Enhanced by machine learning capabilities and AI decision making potential, autonomous military systems have been presented as highly efficient new generation weapons ready to replace conventional armaments and human combatants.

Cyber physical weapon systems, ranging from self-targeting missiles to self-learning ground vehicles are boosting covert military offensive capabilities increasing heated discussions and tensions between nations. The AI cyber arms race, escalated by super powers competing for global dominance, is supported by multi-billion-dollar investments into state-of-the-art technologies and leading research institutions. Further uncontrollable AI-based military technology development will expand the threat landscape with self-replicating cyber weapons and autonomous offense.

The chapter reviews the AI impact on autonomy and its major criteria, explores cyber vulnerabilities in autonomous technologies, highlights critical issues of the AI use in AWS, deliberates on incorporation of ethical principles into development of technologies, reveals legal complications and consequences of AI arms race, forecasts future challenges. It also provides some potential crisis scenarios.

Current research arguments that the generated neural networks and machine learning algorithms, being of complex nature, still remain unpredictable, unreliable and even dangerous when fully autonomous. Joint global action, legally bounding regulations and internationally coordinated research are the major solutions to harness and revert the emerging existential danger.