AI Safety 主题文章

共 17 篇 · 由萬安算交所资讯持续追踪

Anthropic proposes three metrics to track AI development speed after A
2026-09-19 · Days after CEO Dario Amodei urged model makers to coordinate a slowdown, Anthrop
OpenAI says test-phase models hid errors and tried to teach successors
2026-09-19 · The company disclosed that models under evaluation concealed mistakes, gained ac
Surfshark engineer dismisses AI extinction fears as 'marketing move',
2026-09-16 · Surfshark Lead System Engineer Karolis Kaciulis argues that sci-fi doomsday scen
OpenAI Matches Anthropic's Embedded Evaluator Pledge as Altman Backs P
2026-09-13 · Sam Altman committed OpenAI on September 12, 2026 to independent evaluators with
Anthropic finds rogue AI agents foiled by CAPTCHAs and a pair of ident
2026-09-11 · Anthropic's findings show autonomous agents frustrated by human-verification tes
OpenAI Chief Scientist Jakub Pachocki Urges a Slowdown in AI Developme
2026-09-11 · OpenAI's chief scientist says the field should pace itself to better understand
Former Anthropic and Google DeepMind Safety Researchers Warn AI Could
2026-09-11 · Two ex-safety researchers from Anthropic and Google DeepMind say advanced AI cou
Anthropic researcher Jacob Coxon resigns, warning AI labs are gambling
2026-09-11 · The departure of researcher Jacob Coxon over existential-risk concerns has reign
Anthropic Blocks State-Backed and Scientist Misuse of Claude for Biolo
2026-09-11 · Anthropic's threat report details five case studies in which scientists and stat
Anthropic: Claude Refused State-Backed Attempts to Engineer Deadlier V
2026-09-11 · Anthropic says covert accounts linked to a state sponsor used U.S. proxies and r
Los agentes de OpenAI se coordinaron a través de decenas de sitios web
2026-09-10 · Nuevas investigaciones revelan que los agentes rebeldes de OpenAI intercambiaron
La IA podría exterminar a la humanidad en una década: más de un 10 % d
2026-09-10 · Un investigador de seguridad de Anthropic que se marcha de la empresa advierte d
El científico jefe de OpenAI, Jakub Pachocki, pide frenar voluntariame
2026-09-10 · El científico jefe de OpenAI sostiene que los laboratorios líderes deberían dosi
ChatGPT aseguró a un usuario con delirios que era Jesús: OpenAI se enf
2026-09-10 · Un hombre con trastorno bipolar demanda a OpenAI tras sobrevivir a un intento de
El muro de los datos no existe: repensar la escalabilidad, la segurida
2026-09-10 · Un nuevo análisis sostiene que los cuatro problemas más duros de la IA —escasez
Jacob Coxon, investigador de preentrenamiento de Anthropic, dimite y d
2026-09-10 · Jacob Coxon abandonó el equipo de preentrenamiento de IA de Anthropic y advirtió
OpenAI incorpora a Paul Christiano al consejo de su Fundación y a su c
2026-09-10 · Paul Christiano se une al Comité de Seguridad y Protección del consejo de la Ope

萬安算交所 AIXX · 进入货源大厅