🛡️ AI Safety & Alignment

AI safety research, alignment techniques, interpretability, governance, and existential risk discussions.

1,186 articles

What’s the price tag for preventing an AI apocalypse?

The creation of superhuman artificial intelligence could lead to two worst-case scenarios, says Charles Jones, a professor of economics at Stanford Graduate... Source: forbesindia.com

Anthropic Wants You to Know AI Might Kill Everyone – and Also, Please Use Claude

Anthropic's World Cup ad warns AI could end civilization while asking you to trust Claude - but the company's safety record is full of contradictions. Source: tech.yahoo.com

3 Questions: Neural transparency and the future of AI design

MIT Assistant Professor Pat Pataranutaporn describes a new interface that lets everyday users glimpse inside an AI's neural network before their chatbot... Source: news.mit.edu

The Agony of Opacity: Foundations for Reflective Interpretability in AI-Mediated Mental Health Support

Throughout history, a prevailing paradigm in mental healthcare has been one in which distressed people may receive treatment with little understanding... Source: ojs.aaai.org

The small team in Montreal trying to save the world from AI

MONTREAL — Like many of us in early 2023, Yoshua Bengio was tooling around on ChatGPT, a bit in awe of its capabilities. Bengio was at once impressed and... Source: thelogic.co

Google's DeepMind CEO Calls for US-Led Global AI Watchdog

Demis Hassabis proposes new regulatory body with power to halt dangerous AI models. Source: techbuzz.ai

Vatican Hosts Nobel Laureates, Experts To Discuss AI Security Risks

By Ishmael Adibuah. Key Takeaways: Vatican hosts summit on AI and nuclear risks: From July 14–16, 2026, over 200 academics, innovators, and Nobel laureates... Source: eurasiareview.com

AI Tutors Beat Law Professors in Stanford Blind Study, Exposing Bias Risk

As U.S. law schools scramble to set AI policy — the University of Chicago Law School unveiled its strategy just days ago, including plans to ban laptops... Source: techtimes.com

What Anthropic’s latest AI discovery does—and doesn’t—show

The company says it has found a new window into how its models arrive at answers. We spoke with senior editor Will Douglas Heaven about it. Source: technologyreview.com

When the Future Came to City Hall: The Forgotten Blueprint for A.I. Governance

Fifty years after Cambridge's recombinant DNA debate, its model of public oversight offers lessons for A.I. governance. Source: observer.com

COLUMN: Why Explainable AI Is Not Enough for Homeland Security Operations

Homeland security warning rarely arrives as a clean signal. Consider that, at a major public event, cyber defenders may see unusual access attempts,... Source: hstoday.us

How China is ripping off cutting-edge AI from Anthropic, OpenAI — and threatening US national security

AI giants like Anthropic and OpenAI have accused China of ripping off their technology with illicit “distillation” attacks – an alarming trend that has... Source: nypost.com

Hundreds of economists say 'we must act now' on AI’s economic impact and job displacement risks

Hundreds of economists urge immediate action to address AI's potential impact on the economy. In an open letter released Monday, they warn that AI could... Source: timeswv.com

6 months to live for open models

The most serious test to date of open source AI's viability is happening right now. Source: interconnects.ai

Technology Committees: From Optional to Essential

As AI, cybersecurity and emerging technologies reshape risk and value creation, boards are rethinking whether traditional committee structures can keep... Source: directorsandboards.com

AI Data Centers and the Concentration of Wealth

This essay was written with Nathan E. Sanders, and originally appeared in The Guardian. Opposition to AI data centers has emerged as a primary theme in US... Source: securityboulevard.com

Vitalik Buterin: The Real AI Danger Isn’t Superintelligence, It’s Who Controls It

Ethereum co-founder Vitalik Buterin has shifted the AI safety debate, arguing the gravest risk is not superintelligent machines but the concentration… Source: finance.biggo.com

Opinion | India's Opening: What New Delhi Can Actually Do In UN's AI Dialogue

The inaugural Global Dialogue on AI Governance was held in Geneva on 6-7 July and India was represented by the Minister of State for External Affairs Kirti... Source: ndtv.com

Why Attribution Stability Matters More Than Attribution Accuracy

SHAP attribution accuracy is the wrong metric for regulated AI. σ_SHAP — variance across K rotated background samples — is the defensible alternative. Source: hackernoon.com

OpenAI Loses Sixth Safety Leader in Two Years, Folds Team Into Research

OpenAI safety leadership restructuring: head of safety Johannes Heidecke departs July 24 as OpenAI folds its safety teams under research VP Mia Glaese,... Source: techtimes.com