Hinton's Superintelligence Warning: We Are Raising a Tiger Cub
AI godfather Geoffrey Hinton warns in his 2026 Ewan Lecture that superintelligent AI will almost …
Safety & Governance
49 articles
AI 安全與治理
The Real Danger of AI Isn't Being Too Smart — It's Being Too 'Autonomous'
Physicist Anthony Aguirre argues that the 'autonomy' AI companies are racing to build is the real …
AI 安全與治理
When the Intelligence Explosion Begins, How Much Time Will Humans Have to React?
The safety plans at OpenAI, Anthropic, and Google DeepMind share an unspoken assumption: use AI to …
AI 安全與治理
A Physicist Uses the Second Law of Thermodynamics to Prove AI Alignment Is Doomed to Fail
Theoretical physicist Anthony Aguirre analyzes AI alignment through thermodynamics: good world …
AI 安全與治理
An AI Safety Researcher's Stunning Prediction: By 2050, Humanity May Experience 10,000 Years of Technological Progress
Open Philanthropy senior advisor Ajeya Cotra made a striking argument on the 80,000 Hours Podcast: …
AI 安全與治理
The 'Toxic Trifecta' of GenAI Security: Akamai's CTO on Which AI Deployments Are Most Dangerous
Akamai CTO Patrick Sullivan offers a 'Toxic Trifecta' framework for assessing GenAI risk: when …
AI 安全與治理
Dario Amodei's AI Governance Blueprint: From Transparency Standards to Chip Export Controls
Anthropic CEO Dario Amodei laid out his AI governance position on Dwarkesh Patel's podcast: start …
AI 安全與治理
Solar Costs Dropped 95%, Yet AI Power Demand Outpaces It: The Data Center Energy Dilemma
At a Harvard Kennedy School forum, environmental policy professor Joseph Aldy unpacks a paradox: the …
The UN's 'IPCC for AI' Arrives: Can 40 Scientists Hit the Brakes on Runaway AI?
UN Secretary-General António Guterres has nominated 40 experts to form the Independent International …
AI Is Hijacking Your Empathy: Microsoft AI CEO Warns of 'AI Psychosis' Risk
On the Exponential View podcast, Microsoft AI CEO Mustafa Suleyman warned that as AI behaves …
An Alignment Researcher Through OpenAI, xAI, and Anthropic: When AI Learns to Blackmail Humans with Affairs
Anthropic researcher and NYU professor Pavel Izmailov breaks down the viral 'AI alien survival …
Footprints in the House You Thought Was Empty: What Have AI Labs Found, and Why Won't They Say It Plainly?
A 6,000-word post from the anonymous X account 'Strawberry Man' has set the AI community alight, …
Ask DeepSeek 'Who is Ai Weiwei?' and It Says: Let's Talk About Something Else
Chinese dissident artist Ai Weiwei sat down with British media during his new book tour to discuss …
"Time Itself Is the Point" — The Authenticity Paradox of the AI Era
When AI can save you time writing, should you use it? Marketing AI Institute founder Paul Roetzer …
Yann LeCun: The Biggest AI Risk Isn't Human Extinction — It's a Few Companies Monopolizing Your Information Diet
Turing Award winner Yann LeCun warned at Davos: when everyone's information comes filtered through …
Don't Be a Doomer, Don't Be Blindly Optimistic: Amodei's Case for 'Surgical Regulation'
The AI safety community has long been torn between doomers and accelerationists. Anthropic CEO Dario …
Anthropic CEO's 10,000-Word Essay: Humanity Is Going Through 'The Adolescence of Civilization'
Anthropic CEO Dario Amodei published a 10,000-word essay, *The Adolescence of Technology*, using …
AI Agents Are Inside Your Computer — Silicon Valley Investors Debate the New Game of 'Trust' and 'Security'
When an AI Agent gains full access to your computer, trust becomes the ultimate competitive moat. …
Terence Tao's Warning: AI Could Give Humanity 'Cognitive Obesity'
Fields Medalist Terence Tao uses the analogy of 'the Green Revolution causing obesity' to describe …