ETHICAL AND REGULATORY CHALLENGES OF AI DEPLOYMENT IN AUTONOMOUS WEAPONS SYSTEMS

Abstract

This paper analyses the considerations of ethics and the regulation challenges, as applied to the application of artificial intelligence in autonomous weapons systems, which will be conducted using a mixed-method research methodology. To evaluate the accuracy of AI judgment, the probability of committing an ethical violation, the accumulation of risk, or the effectiveness of the human control in different cases of autonomous interactions, quantitative simulations were conducted. The findings demonstrate that AI systems are ethical in most instances but the more autonomous and the less stable environment the less ethical they become. It was observed that ethical risk was increasing over time, and the principles of proportionality and discrimination were regularly violated in the fully independent situations. Graphic presentations of experimental results reveal that there is a declining level of transparency and also the capacity of people to overrule is reducing, which poses some serious questions about accountability. Qualitative regulatory review indicates that existing international legal systems are weak in alleviating these factual risks, as they do not specify enforceable methods to use to enforce real-time compliance with ethics and post-deployment responsibility. The combination of the results highlights a growing gap between technological capability and ethical control, which reveals the urgent need to reinforce international controls, establish human surveillance as a mandatory requirement, and enforce ethical constraints with proven constraints in AI-controlled autonomous weapon systems.