Chapter
Redirecting AI Research Towards Beneficial Outcomes
1:41:42 - 1:50:52 (09:09)
AI researchers are looking to redirect their research towards helpful outcomes, but mid-level AI managers in corporations may not be as motivated to change their approaches to AI development to prioritize beneficial outcomes. There is a need for traditional AI systems to be improved and reevaluated.
Clips
The speaker discusses the possible risks and responsibilities of using AI in the near future, including those that may arise from self-driving cars and the ongoing COVID-19 pandemic.
1:41:42 - 1:42:57 (01:15)
Summary
The speaker discusses the possible risks and responsibilities of using AI in the near future, including those that may arise from self-driving cars and the ongoing COVID-19 pandemic.
ChapterRedirecting AI Research Towards Beneficial Outcomes
Episode#364 - Stuart Russell - The Terrifying Problem Of AI Control
PodcastModern Wisdom
The constant progress of technology and our attachment to it could lead to people ignoring the potential risks and threats posed in biology as well as DNA synthesis.
1:42:58 - 1:44:46 (01:48)
Summary
The constant progress of technology and our attachment to it could lead to people ignoring the potential risks and threats posed in biology as well as DNA synthesis. The building of black box circuitry by manufacturers of DNA synthesis equipment is worrying as it checks the synthesis of circuits nonpublicly, and similar things may be happening in biology.
ChapterRedirecting AI Research Towards Beneficial Outcomes
Episode#364 - Stuart Russell - The Terrifying Problem Of AI Control
PodcastModern Wisdom
The rise of libertarianism in synthetic biology and the push towards decentralization in web 3.0 may lead to a similar approach towards AI and computer science in the future, despite potential risks and consequences.
1:44:47 - 1:47:50 (03:03)
Summary
The rise of libertarianism in synthetic biology and the push towards decentralization in web 3.0 may lead to a similar approach towards AI and computer science in the future, despite potential risks and consequences.
ChapterRedirecting AI Research Towards Beneficial Outcomes
Episode#364 - Stuart Russell - The Terrifying Problem Of AI Control
PodcastModern Wisdom
Despite traditional AI systems being criticized and a lack of innovation, many AI managers may not be incentivized to change their approach and could continue using standard models and reinforcement learning libraries.
1:47:51 - 1:50:52 (03:00)
Summary
Despite traditional AI systems being criticized and a lack of innovation, many AI managers may not be incentivized to change their approach and could continue using standard models and reinforcement learning libraries.