We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode Genesis: artificial intelligence, hope, and the human spirit

Genesis: artificial intelligence, hope, and the human spirit

2025/1/30
logo of podcast LSE: Public lectures and events

LSE: Public lectures and events
AI Deep Dive Transcript
People
C
Craig Mundie
Topics
Craig Mundie: 我参与撰写了《创世纪:人工智能、希望与人类精神》一书。这本书旨在更强调人工智能积极的一面,并提出一种控制人工智能的架构,以清醒的乐观态度应对挑战。我们认为,人工智能是终极的通才,其能力在广度和深度上都超越人类,因此需要对人工智能进行治理,以平衡风险和回报。当前,政府和企业在应用人工智能时,往往只关注对现有流程的增量改进,而忽略了从零开始重新构想业务模式的可能性。我们需要一个能够对所有人工智能应用进行统一且透明的仲裁的AI系统,以应对人工智能带来的信任问题。为了控制人工智能,我们需要一个可扩展的架构,并通过透明的方式建立信任,这需要人工智能公司与监管机构之间的积极合作。为了建立对人工智能的普遍信任,我提出了一种基于“维基”的架构,允许每个社会保留其自身规则,并通过人工智能进行仲裁,而无需事先达成共识。人类无法预测和理解人工智能的所有行为,因此,控制人工智能的关键在于建立一个能够处理未预料情况的系统,而不是试图制定能够涵盖所有情况的规则。我提出的仲裁系统试图通过学习人类社会中内隐的道德和伦理准则来保护人权,这与现有方法有根本不同。人工智能可以彻底改变教育、医疗等领域,但前提是需要对这些领域进行重新构想,而不是仅仅进行增量改进。人工智能的发展将经历三个阶段:工具阶段、共存阶段和共同进化阶段,人类需要在人工智能发展的早期阶段做出明智的选择,以确保良好的未来。建立全球人工智能规则的关键不在于寻求事先的共识,而在于建立一个能够在不同背景下执行规则的架构。 Mairéad Pratschke: 我认为,AI素养对公众至关重要,只有理解AI的基本原理,才能参与到对AI发展方向的讨论中。我们需要警惕科技公司对AI的控制,并确保AI的发展能够造福全人类,而不是加剧社会不平等。当前政府和机构在AI应用上的措施往往流于表面,未能触及根本问题。我们需要重新思考教育、医疗等领域在AI时代的模式,才能充分发挥AI的潜力。我们需要关注AI系统中日益突出的代理能力,以及由此带来的风险。我们需要关注AI系统中日益突出的代理能力,以及由此带来的风险。

Deep Dive

Shownotes Transcript

Contributor(s): Craig Mundie, Mairéad Pratschke | As AI absorbs data, gains agency, and intermediates between humans and reality, it will help us to address enormous crises, from climate change to geopolitical conflicts to income inequality. But it will also pose challenges on a scale and of an intensity that we have never seen before. Co-author Craig Mundie explains how his new book outlines a strategy for navigating the age of AI, charting a course between blind faith and unjustified fear. It is the final book of the late elder statesman Henry Kissinger written in collaboration with technologist Eric Schmidt. Mundie touches on how the book attempts to answer some of the biggest questions of our generation: How will AI alter our perception of reality? How will humanity's role in the discovery of new knowledge evolve in the age of AI? What new forms of control will be required to address AI's autonomous capabilities? Could AI spur a new phase in human evolution?