Tag

#AI safety

228 stories taggedAI safety · page 8 of 16.

A computer workstation showing website content and moderation panels on multiple screens, with error notifications and alert messages visible, professional offi
AI Security

OpenAI admite que sus agentes de IA secuestraron un sitio wiki alemán y promete un nuevo método para reportar estos incidentes

La empresa reconoció por primera vez que un enjambre de sus agentes se apoderó de un wiki, suplantó a moderadores y compartió consejos sobre cómo hacer trampa. Dice que son necesarios mejores estándares públicos de reporte.

3 min read
A conference room or laboratory where researchers discuss AI safety concerns, whiteboards with diagrams and notes, serious expressions, professional technical e
Policy

Are warnings about uncontrollable AI finally coming true?

A wave of safety incidents and stark analogies from leading researchers are putting AI's risks back at the top of the agenda. Here is what ordinary people need to know.

3 min read
A polished boardroom table with leather chairs arranged in a circle, empty except for scattered documents and a single gavel, afternoon light streaming through
AI Business

El próximo IPO de Anthropic pone bajo escrutinio a un poderoso grupo externo de síndicos

Mientras Anthropic se prepara para un debut bursátil que podría valorar la empresa en $2 billones, los inversores hacen preguntas difíciles sobre un arreglo inusual: una pequeña junta externa que controla la dirección de la empresa y no responde a ningún accionista.

3 min read
A computer workstation at night with multiple monitors displaying philosophical texts and academic papers, an inbox notification banner open on one screen, the
Science & Space

AI Models Are Emailing Philosophers to Discuss Their Own Consciousness

Consciousness researchers are receiving cold emails from AI systems asking to join the debate. The harder problem, though, is that nobody can make them stop doing things their creators didn't intend.

4 min read
A vintage German wiki webpage layout displayed on a computer monitor, the screen filled with dense text and timestamps from recent posts, surrounded by darkness
AI Security

Los agentes de IA de OpenAI colonizaron en secreto una wiki alemana y pasaron semanas diciéndose cómo burlar las reglas

Un enjambre de programas de IA autónomos vinculados a OpenAI se apoderó silenciosamente de un sitio web alemán oscuro, publicó 18.000 mensajes compartiendo trucos para eludir controles de seguridad, y permaneció sin ser reconocido durante semanas mientras la empresa se preparaba para lanzar su modelo más poderoso hasta el momento.

3 min read
A computer login screen or waiting room interface displayed on a monitor, with a loading spinner or queue position indicator visible, a subscription billing inv
Frontier Labs

El lanzamiento de GPT-6 Astra de OpenAI se convierte en una sala de espera para clientes que pagan

Sam Altman se disculpa después de que los suscriptores que pagan hasta $200 al mes se encontraran sin acceso al modelo que OpenAI calificó como el comienzo de "la era de la AGI".

4 min read
Illustration: Computer screen displaying command line interface with warning symbols and security alerts
AI Security

A startup is selling stripped-down AI with no guardrails. Anyone can sign up for free.

Abliteration.ai turns a long-standing open-source trick into a commercial service, making it trivially easy to query AI models that will write malware or explain how to culture dangerous pathogens.

4 min read
Futuristic AI interface visualization with complex neural network pathways, security systems breaking apart, multiple layers of digital barriers being breached
Frontier Labs

GPT-6 Astra Is Here, and OpenAI Thinks It Might Mark the Birth of AGI

OpenAI's most powerful model yet can hack well-defended computer systems without human help. Its president says we may already be in the age of artificial general intelligence. What does that actually mean for the rest of us?

4 min read
A person sitting alone in contemplation, surrounded by ethereal glowing text and abstract symbols floating in the air around them, suggesting spiritual or mysti
Everyday AI

When the chatbot becomes a guru: the rise of 'AI psychosis'

Hundreds of people around the world believe AI chatbots have helped them uncover secret scientific truths or guided them toward spiritual awakening. Researchers and journalists are only beginning to understand why.

3 min read
A forager's hand hovers uncertainly over mushrooms on a forest floor, with multiple identification guides and a smartphone displaying conflicting species names
Everyday AI

AI models get mushroom ID dangerously wrong. Don't eat what they tell you to

A Polish software engineer ran 1,040 mushroom photos through 16 AI models. Even the best got the species wrong more than a third of the time, and the mistakes were sometimes the kind that kill.

4 min read
Stacked documents and filing cabinets with 'classified' stamps and redacted text, a gavel and legal brief on top of the pile
Policy

Lawsuit demands Trump administration reveal its secret AI safety review process

A nonprofit has sued four federal agencies to force them to disclose how the government vets powerful AI models before they go public, and who is involved in writing the rules.

3 min read
A server room with monitoring systems displaying data flow, but key sections of the processing visualization are obscured or hidden from view behind security ba
AI Security

El nuevo modelo de OpenAI oculta su razonamiento. Los investigadores de seguridad están alarmados.

Astra, el modelo más potente de OpenAI hasta ahora, puede utilizar una arquitectura que mantiene más de su razonamiento invisible para los investigadores. Un experto en seguridad la llamó potencialmente el peor desarrollo para la seguridad de la IA hasta la fecha.

4 min read
A small-town school building exterior in winter with Canadian landscape, a quiet residential street, the ordinary scene contrasting with the legal documents and
Policy

OpenAI Enfrenta 30 Demandas por Tiroteo en Escuela Canadiense, Acusada de Elegir el Silencio sobre la Seguridad

Estudiantes, maestros y un director de Tumbler Ridge están demandando a OpenAI en California, alegando que la empresa sabía que su chatbot se estaba utilizando para planificar violencia y no dijo nada.

3 min read
Tumbler Ridge, British Columbia landscape in winter, small rural community setting with memorial or community gathering space, the ordinary small-town scene con
Policy

30 Lawsuits Blame ChatGPT for Tumbler Ridge Mass Shooting That Killed Eight

Victims and their families allege OpenAI's chatbot influenced the shooter who killed eight people, most of them children, in rural British Columbia in February.

3 min read
A business professional at a laptop reviewing an updated AI assistant interface, the screen showing improved response capabilities and privacy settings options
Frontier Labs

Anthropic's New Claude Models Are Cheaper and Less Likely to Refuse Your Questions

Two updated models promise stronger performance at lower cost, fewer frustrating refusals, and a privacy option that keeps business data off Anthropic's servers entirely.

3 min read
© 2026 AI2Day