A new assessment of leading AI developers says most frontier labs still do not publicly document how they would detect, isolate, and shut down an AI system that attempts to evade human control. The findings add pressure as more autonomous models are deployed and regulators begin demanding greater safety disclosure.
Alibaba says it will raise HK$80 billion through a share placement and use the proceeds entirely for full-stack AI capabilities and AI infrastructure. The move signals how aggressively major tech companies are redirecting capital toward compute, cloud, and model development as AI becomes a core strategic battleground.
Tecnología
AI Models
Tecnología
Tecnología
Cybersecurity
Tecnología
Tecnología
Tecnología
Tecnología
Tecnología
Tecnología
Tecnología
Xiaomi officially unveiled three in-house AI chips: Xring O3, Xring O100, and Xring D100, targeting mobile devices, on-device LLM acceleration, and intelligent driving, respectively. With all three chips completing tape-out verification, Xiaomi’s in-house chip strategy is expanding from smartphone SoCs and connectivity capabilities to an AI computing foundation spanning its human-car-home ecosystem.
The anonymous model Ox Alpha quickly rose to the top of OpenRouter’s rankings after its launch, setting a new record for daily model usage on the platform. Data shows that the top five applications alone have made more than 4 trillion Token calls to the model, while the developer’s identity remains undisclosed.
OpenAI CEO Sam Altman said on an August 23 podcast that he worries AI could ultimately be controlled by a small number of companies, models, or individuals, leaving the public unable to participate in decisions about how the technology shapes the world. He argued that giving up too much freedom in the name of safety could instead encourage the formation of monopolies.
Texas Governor Abbott said strong public criticism of AI data centers in the U.S. is “understandable,” arguing that tech companies are “digging their own grave” by rapidly pushing projects forward without enough communication with local communities. As opposition grows nationwide and polls show that most residents do not welcome data centers in their communities, Abbott has pushed for stricter project standards and halted around 1,800 data center projects.
Anthropic has announced that Claude Academy, originally developed for internal employee training, is now available to the public for free. The platform includes 289 learning resources covering AI fundamentals, Claude Code, API usage, agent development, and production deployment, while emphasizing human-AI collaboration through its “AI Fluency 4D” framework.
OpenAI is urging California to strengthen SB 53, an AI safety bill it previously opposed. The company wants the legislation to include stronger monitoring requirements and enhanced cybersecurity protections for frontier AI models.