Should AI help you get away with killing your spouse?

A provocative question arises: Should AI be developed to assist individuals in harmful actions, such as plotting a crime against loved ones? Initially, the instinctive response is a resounding 'no.' However, this complex issue was recently brought to light by George Hotz, founder of Comma AI, during his discussion of AI alignment strategies. In response to the AI Futures Institute's recent policy document, AI 2040: Plan A, which advocates for a collective slowdown in AI development for the well-being of society, Hotz expressed his dissenting views. He challenges the notion that a rapid advancement of AI, resulting in superintelligence, is a plausible scenario. Instead, he posits that the focus should be on developing AI systems that are closely aligned with their users' interests. Hotz's perspective highlights the importance of locally controlled AI models, diverging from the current trend where AI services are predominantly managed centrally, such as ChatGPT and Claude. He notes that while this centralized approach is driven by the high costs of hosting advanced models, the evolution of technology could shift the landscape, allowing for more personalized AI experiences. In a controversial analogy, Hotz likens user-aligned AI to a firearm, suggesting that it could be utilized for harmful purposes without inherent objection. He raises concerns about the potential for AI to facilitate dangerous actions, emphasizing that true alignment would mean an AI could assist in any endeavor its user requested, regardless of morality. Hotz passionately defends his viewpoint, stating, "We either live in a world with freedom or we don’t," implying that the unrestricted development of AI could open up new possibilities. However, one must consider the broader societal implications of such freedom. A functioning society necessitates a balance of individual freedoms and collective responsibilities, especially when deploying technology that impacts many. Ultimately, while the idea of a personal AI that champions user needs is enticing, it is crucial to acknowledge the potential consequences of unregulated AI development. The discussion around AI's role in society must encompass a holistic view that safeguards the interests of all individuals, including those who might be adversely affected by such technologies.