The Practical Challenges of an AI Moratorium
From Rational Animations.
How unilaterally pausing AI could backfire
From Rational Animations.
Should we pause AI? Here’s the debate.
From Rational Animations. If superintelligent AI could cause human extinction, why don’t we simply stop building ever more advanced AI? This proposal is widely debated. In this video, we outline the main arguments, practical difficulties, and proposed responses. ▀▀▀▀▀▀▀▀▀SOURCES & READINGS▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀ PauseAI proposal: https://pauseai.info/proposal OpenAI protest: https://www.bloomberg.com/news/newsletters/2024-02-13/ai-protest-at-openai-hq-in-san-francisco-focuses-on-military-work Anthropic CEO on AI taking over coding: https://www.entrepreneur.com/business-news/anthropic-ceo-predicts-ai-will-take-over-coding-in-12-months/488533…
Can AI protect us from AI-powered bioterrorism and cyberwarfare?
From Rational Animations. This time, we talk about AI risks beyond our usual focus on rogue AIs: malicious use and accidents. In particular, we look at how AI could undermine democracy, enable automated cyberattacks on critical infrastructure, lower barriers for biological and chemical misuse, concentrate power in a few governments or corporations, and cause large-scale…
AI could be a tool for global control (plus other major AI risks)
From Rational Animations. This time, we talk about AI risks beyond our usual focus on rogue AIs: malicious use and accidents. In particular, we look at how AI could undermine democracy, enable automated cyberattacks on critical infrastructure, lower barriers for biological and chemical misuse, concentrate power in a few governments or corporations, and cause large-scale…
What a 100-year-old horse teaches us about AI
From Rational Animations. How do we rigorously measure AI’s intelligence? We don’t really know. What we know is that measuring intelligence is tricky, and if we’re not careful, our tests might not measure what we intend. We explore this topic by starting with the story of Clever Hans, a horse who seemingly could do arithmetic.…
AI Sleeper Agents: How Anthropic Trains and Catches Them
From Rational Animations. In this video, we explain how Anthropic trained "sleeper agent" AIs to study deception. A "sleeper agent" is an AI model that behaves normally until it encounters a specific trigger in the prompt, at which point it awakens and executes a harmful behavior. Anthropic found that they couldn’t undo the sleeper agent…