AI’s rapid advancement has led to concerns about existential risks it poses, including unintended consequences and intentional misuse by malicious entities. The discourse surrounding AI risks often focuses on “Terminator scenarios” while neglecting the more pressing threat from bad actors intentionally weaponizing AI. Unintended consequences refer to scenarios where AI systems inadvertently cause catastrophic harm as they operate outside human control. Companies have strong incentives to avoid harming customers and engage in pre-deployment testing to correct problems. However, the threat of bad actors weaponizing AI has been largely overlooked in current discourse.
Foreign adversaries and radical extremists seek to gain a technological advantage through AI, posing a real concern for national security. Anti-terrorism policies may provide a roadmap for dealing with these threats, requiring involvement from multiple levels of government. AI safety legislation, such as California’s draft AI bill, focuses on regulating large technology companies rather than addressing potential misuse by bad actors. This misplaced focus raises questions about the true motives of proponents and may fail to address actual risks posed by AI falling into the wrong hands.
To effectively mitigate AI risks, the national security establishment must prepare for technological advancements and focus on the most pressing threats posed by bad actors. While tech companies have a responsibility to develop safeguards against hackers, governments cannot simply dictate terms to big businesses and expect AI safety to be ensured. Existential risks from AI are more likely to come from bad actors, both state and non-state, rather than the few large technology companies dominant in Silicon Valley. It is crucial to broaden the policy response to address the most pressing threats to national security and prevent catastrophic harm.
