The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) in the United Arab Emirates unveiled a new AI world model and agent, two large language models (LLMs), and a research center in Silicon Valley as part of the country’s growing investment in the field. The president and professor of MBZUAI, Eric Xing, introduced the models and lab at the Computer History Museum in Mountain View, California. The UAE’s interest in AI has been on the rise, thanks to the guidance of Sheikh Tahnoun bin Zayed al Nahyan, the tech-savvy national security adviser and brother of President Mohamed bin Zayed Al Nahyan.

The new AI world model, PAN, is designed to create physically realistic simulations for testing AI agents. This technology will help in refining the performance of AI agents in various scenarios. Additionally, MBZUAI launched PAN-Agent, an experimental AI agent trained to handle reasoning tasks within the PAN world model. The aim is to simulate real-world situations, such as self-driving cars on virtual roads, using AI technology. The unveiling event showcased PAN’s capabilities in simulating self-driving cars navigating through busy roads, drones flying in unfamiliar spaces, and robots operating in domestic environments.

The UAE’s investment in AI extends to the development of two new large language models (LLMs) – K2 and Jais. K2, a 65-billion parameter model optimized for reasoning tasks, was trained on 80 A100 chips using Nvidia’s DGX Cloud. MBZUAI claims that K2 was developed using 35 percent less compute power compared to Meta’s Llama 2 at the same size. The university also introduced Jais, touted as the world’s most advanced Arabic-language LLM. These advancements in AI technology are crucial for building more advanced AI systems like virtual assistants and robots capable of functioning in unfamiliar environments. The unveiling of these models and the establishment of a new research center in Silicon Valley signify the UAE’s commitment to tapping into global AI knowledge and talent to drive innovation in the field.