How Saudi Arabia's KAUST is building the tools, talent, and vision for generative AI

Arab News

2 days ago

DHAHRAN: As Saudi Arabia accelerates its push to become a global hub for artificial intelligence, the research labs at the King Abdullah University of Science and Technology are working to ensure the Kingdom doesn't just consume AI — but helps build it.
On KAUST's Red Sea campus, the newly established Center of Excellence in Generative AI — CoE Gen AI — is at the heart of this effort.
One of its founding members, Prof. Peter Wonka, is leading a team developing foundational models and tailored tools designed to align with Saudi Arabia's national priorities — from personalized education and energy modeling to AI-generated Arabic content.
'This is a really exciting time to be involved in AI,' Wonka, who is originally from Austria, told Arab News. 'This is the time of tremendous progress.'
Saudi Arabia's ambitions are part of a global race to dominate generative AI. PwC estimates that AI will contribute about SR 878 billion ($235 billion), or roughly 12.4 percent of Saudi Arabia's gross domestic product, by 2030.
According to research from McKinsey, generative AI could add between $2.6 trillion and $4.4 trillion annually to the global economy — with around 75 percent of that impact concentrated in customer operations, marketing, software engineering, and research and development.
Under its national AI strategy, Saudi Arabia has pledged approximately $20 billion in cumulative investments by 2030, aimed at building AI infrastructure, research, and talent development, according to Reuters.
'It has been a very exciting year discussing with various industries in Saudi Arabia about their AI adoption and possible projects we can work on together,' said Wonka. 'Our 30 faculty cover a lot of expertise, making us an excellent and reliable partner for many industries in the Kingdom.'
The CoE Gen AI currently includes faculty members working across five major research initiatives, supported by KAUST's Visual Computing Center, where Wonka serves as associate director.
Each project maps to one of Saudi Arabia's national research, development, and innovation priorities — health and wellness, sustainability and essential needs, energy and industrial leadership, and economies of the future — alongside foundational AI research that supports applications across domains.
One of KAUST's major translational focuses is education.
Earlier this year, Arab News reported that AI would be integrated into the school curriculum starting in the coming academic year. The introduction of this nationwide AI curriculum aims to support the Kingdom's Human Capability Development Program, part of the Vision 2030 plan for national development and diversification.
The curriculum was unveiled in July by the National Curriculum Center, with the Ministry of Education, Ministry of Communications and Information Technology, and the Saudi Data and Artificial Intelligence Authority. It features age-appropriate AI modules in the form of interactive and hands-on teaching.
But Wonka cautions that simply adding AI isn't enough. 'It only makes sense with an additional understanding of other subjects and with a good grasp of fundamentals,' he said.
At the university level, AI tools also have the potential of being misused. The result is a widening gap between students who treat AI as a shortcut and those who strategically integrate it into a broader learning workflow.
'People that possibly grew up without AI have better fact-checking skills,' said Wonka, emphasizing that over-reliance on these tools can produce incorrect or low-quality outputs.
He encourages students to use generative tools to brainstorm ideas, rewrite for clarity, or structure outlines — but warns against letting the model replace original research. 'A paper written entirely by AI has no research, no ideas,' he said.
That tension between power and reliability is central to KAUST's AI safety research. 'These tools can give wrong answers very confidently and sometimes even mix up elementary facts,' he said.
'Still, it may be too optimistic to hope for AI tools that give no wrong answers. A more pragmatic approach would be to learn how to use AI tools despite the fact that they may give wrong answers from time to time.'
Beyond education, generative AI is poised to impact nearly every sector. 'For business communication, AI is already used for spell-checking, editing, and drafting complete letters or emails,' said Wonka.
'Doctors will heavily lean on AI tools to support diagnosing their patients … Cars will have more and more automatic driving features to assist humans in driving. Creative work, such as graphic design and photography, will greatly rely on AI assistance.'
Yet, while the technology is advancing quickly, Wonka remains cautious. 'That is not something I would feel very comfortable predicting,' he said. 'A lot of people are just speculating.'
Instead, he envisions a more subtle shift — AI gradually embedding itself into daily workflows. 'What I'm betting on happening almost certainly is more that AI will be integrated in everybody's life to make it 25 percent more efficient.'
A key part of the CoE Gen AI's mission is to build tools that are culturally relevant and locally deployable.
'There are a lot of different companies that are interested in having local use of AI so that they don't send all their data to the big companies — because they're really worried about that,' said Wonka.
To that end, KAUST is also investing in Arabic-language AI systems for tutoring, content generation, and search.
'The main emphasis of our work in AI in intelligent tutoring is to develop tools for personalized learning, personalized testing, and dual language support in both English and Arabic,' said Wonka.
In visual content generation, his team is focused on personalization and the creation of culturally relevant content in both languages.
Meanwhile, in the realm of smart cities, KAUST researchers are exploring how AI can support digital twins, architectural data analysis, and urban planning. 'Cognitive cities are able to learn, adapt, predict, and proactively respond,' he said.
One open question, he added, is how far AI will go in replacing human roles altogether.
KAUST has launched a new Center of Excellence in Generative AI to develop foundational models aligned with national priorities.
The research spans five core areas, including personalized education, sustainable energy modeling, Arabic content generation, and smart city planning.
Prof. Peter Wonka leads the initiative, emphasizing both academic rigor and practical industry collaboration within the Kingdom.
'Will we see most doctors' offices staffed only by robots, or will there only be human doctors using AI tools? It's important for everyone to invest time in learning about AI and using AI tools.'
Success, he said, is measured both academically and practically. 'For translational impact, the university considers metrics such as the success of startup companies, the number and size of in-Kingdom collaborations and their impact, patents and their generated revenue, as well as successful research collaborations within the Kingdom and abroad that are externally funded.
'Ultimately, a lot of the success of the university is about collaborations. The CoE Gen AI aims to support in-Kingdom businesses and government entities to develop impactful AI projects.'
Wonka joined KAUST in 2012 after stints at Arizona State University and Georgia Tech. With a background in computer science, computer vision, and urban modeling, he has helped shape KAUST's AI vision from its early stages to its current, ambitious scope.
Now, as one of the Kingdom's leading researchers in generative AI, he is focused on ensuring that Saudi Arabia, his home for the last 14 years, is not just reacting to the global AI wave — but actively shaping it.
'I do believe that you'll see a lot of impact coming from AI,' he said. 'And I think as a researcher, it's really exciting to be on the forefront of something that is that exciting.'