Artificial Intelligence in Defense has huge potential - it can be trained to generate alerts for threats by recognizing abnormal behavior, identify new types of malware, and protect data for security organizations. Put in others' hands, it can also open the possibility for individuals to launch cyber attacks. AI, for instance, can use algorithms to crack systems and break into networks, and launch attacks that are harder to defend against spontaneously due to the frequency that they can be produced. Is it possible for safeguards to be put in place to prevent the unethical manipulation of AI systems? Do the potential costs of AI’s misuse outweigh its potential for national security?