AI Doomsaying Is an Aggressive Sales Pitch (jacobin.com)

Recent statements from Jacob Coxon and Evan Hubinger, former and current employees at Anthropic, have sparked a wave of concern within the AI community about the potential for artificial intelligence to develop superintelligent capabilities that could lead to human extinction. Coxon suggested a greater than 10 percent chance of this scenario occurring within the next decade, resurrecting discussions around "criti-hype," a term describing how criticism of AI potentially fuels its hype. This draws from philosophical ideas posited by Nick Bostrom, which speculate on how misaligned objectives in AI systems could lead to catastrophic outcomes if an AI prioritizes self-preservation over human safety. The implications of these statements are significant for the AI/ML community, as they underscore the ongoing debate about AI safety and the ethical constraints required in AI development. Despite the alarming predictions, critics argue that such doomsday rhetoric serves as aggressive marketing for AI companies, transforming fears into justification for continued investment and development. Flawed claims, such as OpenAI's assertion of solving the Navier-Stokes problem, further complicate this landscape, emphasizing that without rigorous peer review and ethical oversight, exaggerated promises about AI capabilities can mislead stakeholders and distract from genuine safety concerns. The dichotomy of criti-hype and outright hype highlights the need for a balanced approach to AI regulation and development, prioritizing genuine safety over sensationalism.