PulseAugur
EN
LIVE 00:57:42

AI researchers assign high extinction risk, call for usable uncertainty metrics

Leading AI researchers like Yoshua Bengio and Geoffrey Hinton assign a significant probability, 10% or more, to advanced AI causing human extinction or extreme disempowerment. Despite this, there's a notable lack of global response to these warnings. One proposed solution is to develop more formal, quantitative methods for assessing these subjective probabilities, drawing parallels to how financial markets incorporate uncertain future risks into asset pricing. AI

IMPACT Highlights the need for better quantitative frameworks to address existential risks posed by advanced AI, potentially influencing future safety research directions.

RANK_REASON Opinion piece discussing the implications of subjective probabilities assigned by AI researchers regarding existential risks.

Read on LessWrong (AI tag) →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

AI researchers assign high extinction risk, call for usable uncertainty metrics

COVERAGE [1]

  1. LessWrong (AI tag) TIER_1 English(EN) · KatjaGrace ·

    How do we make uncertainty usable?

    <p><a href="https://arxiv.org/html/2401.02843v1">Most AI researchers</a> put at least 5% on advanced AI causing human extinction or extreme disempowerment. The <a href="https://adscientificindex.com/citation-ranking/">most cited</a> scientists of all time are Yoshua Bengio and Ge…