Google has accelerated the release of its Gemini AI models, but has not yet published safety reports for these models. This has raised concerns about the company prioritizing speed over transparency. While Google claims that safety remains a top priority and plans to release more documentation, the lack of model cards for these recent releases…
Google DeepMind is exploring the frontiers of artificial general intelligence (AGI) with a focus on readiness, proactive risk assessment, and collaboration with the AI community. The article emphasizes the importance of mitigating safety risks associated with AGI, such as misuse and misalignment, through systematic and comprehensive approaches, and highlights the need for ongoing dialogue within…
The article discusses China’s ambitious efforts to revolutionize the use of humanoid robots in the workplace, aiming to overcome current limitations that make these robots inefficient co-workers. By investing heavily in advanced AI and robotics technologies, China hopes to lead the way in creating humanoid robots that can seamlessly integrate into various labor-intensive industries, potentially…
Security researchers discovered a significant vulnerability in Unitree Go1 robot dogs. The flaw involves an undocumented remote access tunnel service called CloudSail (Zhexi), allowing unauthorized individuals to control the robots and access their cameras using default credentials and a specific API key. This poses serious risks, especially in sensitive environments, as the robots are used…