For AI leaders Doom is a form of hype (erkansaka.net)

A recent statement by Evan Hubinger, head of alignment science at Anthropic, claiming a "greater than 10% chance" of AI causing human extinction within a decade, has sparked concern in the tech community. This alarming assertion coincided with the resignation of Jacob Coxon, a former pretraining researcher at OpenAI and Anthropic, who criticized both companies for irresponsibly racing toward self-improving superintelligence. This dialogue highlights a significant tension within the AI community, as senior figures grapple with the implications of their innovations while also advocating for regulatory frameworks that may benefit their organizations. The discourse surrounding AI's potential dangers has become a form of hype that may influence public perception and policy-making. Critics argue that these doomsday narratives serve dual purposes: fostering fear to justify regulatory exemptions for leading companies while simultaneously stifling meaningful discussions about the tangible harms already caused by AI, particularly to low-wage workers and marginalized communities. The AI/ML community faces an urgent need for a more nuanced conversation that balances innovation with ethical regulation, focusing on the real and pressing challenges rather than speculative apocalyptic predictions.