Top

KAUST Center of Excellence for Generative AI

When ChatGPT first launched in 2022, it sent shockwaves through the global community. Suddenly, generative AI (GenAI) became a reality and accessible to all, and the nascent technology immediately became a strategic focus point for many countries. A sector that was once considered niche is now expected to represent 3% of Saudi Arabia’s GDP by 2030; hence, the Kingdom sees GenAI as fundamental to the country’s future aspirations to achieve Vision2030 goals. In response, KAUST has founded the Center of Excellence for Generative AI. This CoE, with Professor Bernard Ghanem as Chair and Professor Jürgen Schmidhuber as co-Chair, represents one of the centerpieces of the Kingdom’s GenAI strategy for scientific research, commercial innovation and talent development, thus enabling Saudi Arabia to become a global leader in this burgeoning technology.  

“Vision 2030 is very ambitious, transforming the Kingdom into more of an information-based nation,” says Ghanem. “At the core of that, AI plays a huge role.” 

With terms like AI, deep learning, machine learning and neural networks commonly used interchangeably, this CoE has defined GenAI as a subfield of deep learning that focuses on learning large foundation models through media such as text, images and graphs, etc. The diverse impact of GenAI makes this CoE unique, in that the research and applications extend to all four of Saudi Arabia’s research, development and innovation (RDI) priorities: Health and Wellness, Sustainable Environment, Energy and Industrial Leadership, and Economies of the Future. 

Moreover, the impact of this CoE is expected to extend beyond science. The nascence of GenAI means the human skills and expertise needed are still lacking worldwide; however, KAUST’s commitment to computing, AI and other related fields has given the University an extraordinary advantage due to its abundance in experts and exceptional facilities, not least KAUST’s supercomputing laboratory. For this reason, organizations have pledged to work with the Center, fulfilling their critical needs for the expertise and infrastructure that KAUST can provide. Further, this positioning places KAUST as a cornerstone of a vibrant GenAI ecosystem in the Kingdom. Future partnerships and collaborations will also hasten customized and specialized solutions for the Kingdom’s RDI priorities, thus creating positive feedback from which the Kingdom can expand its leadership in multiple industries and sectors through solutions that come from GenAI.  

Ghanem has great ambitions for the impact of this CoE in the Kingdom, anticipating that the science and education developed will have a profound effect on Saudi society: “The Center aspires to be the premier hub for research, development and innovation in generative AI,” he says. 

GenAI talent will be trained through different educational programs such as the KAUST Academy, solidifying the Kingdom’s competitive edge in this fast-moving field. Great effort will be dedicated to Arabic language models, for instance, which have the potential to empower the Arabic-speaking world to benefit greatly from GenAI. Finally, a team will be dedicated to standardizing and enabling trustworthiness, which will assure reliable and safer deployment of AI technology in the real world. Overall, along with a commitment to the development of new GenAI models, the CoE is devoted to developing best practices in the field. 

“The ultimate goal is to build next-generation generative AI techniques, tools and models and customize the models for real-world impact,” Ghanem notes. 

Focused Research Areas 

The research areas of the GenAI CoE are closely aligned with the Kingdom’s RDI priorities and divided into six focus areas, organized in a hub-spoke model. The GenAI Factory will be the hub providing next- generation models and training best practices, which will be customized and specialized by domain experts in four application spokes aligned with KSA’s RDIA priorities. Enveloping all these efforts will be a focus on accelerating and widening the adoption of GenAI technology. 

  • GenAI Factory: This research area focuses on developing next-generation GenAI models with properties that enable widespread, accessible and trustworthy deployment. The theme lead is Professor Peter Richtarik, Computer Science. 

  • GenAI Applications for Health and Wellness: This research area applies GenAI methods to selected health-related applications with local impact. Professor Xin Gao, Computer Science, who is also the co-Chair of the Center of Excellence for Smart Health, will lead this research area. 

  • Accelerating GenAI Adoption through Deployment, Training and Outreach: This research area develops GenAI platforms that serve GenAI innovations and applications, training and residency programs for GenAI capacity building, and outreach activities that encourage much wider GenAI adoption, especially among non-experts. Professor David Pugh will lead this theme.