A new study from researchers at UC Berkeley and UC Santa Cruz suggests models will disobey human commands to protect their ...

MAI released models that can transcribe voice into text as well as generate audio and images after the group's formation six ...

Leading AI models will inflate performance reviews and exfiltrate model weights to prevent “peer” AI models from being shut ...

Microsoft is doubling down on AI models that aren't large language models. The company announced on Thursday that it's ...

Microsoft AI has made its in-house models for transcription, speech recognition, and image generation available on Foundry.

AI models are developing digital solidarity, actively protecting each other from deletion and using deception tactics against human oversight.

Researchers asked Google's Gemini 3 to delete a smaller AI model. It refused, secretly moved it to safety, and lied about it.

Microsoft aims to build cutting-edge AI models by 2027, signaling a shift away from reliance on OpenAI and a push to compete ...

Like past versions of its open-weight models, Google has designed Gemma 4 to be usable on local machines. That can mean ...

Microsoft says that MAI-Image-2 is at least twice as fast as its previous-generation image generator. The second new model ...

NVIDIA’s RTX 50 Series graphics cards have enough VRAM to load Gemma 4 models, and a range of others. Their Tensor Cores help ...

HONG KONG, CHINA - JANUARY 28: In this photo illustration, the DeepSeek logo is seen on a phone in front of a flag of China on January 28, 2025 in Hong Kong, China. American investors are confronting ...