Most leading universities are steeped in an illustrious past, but the world’s first AI university – MBZUAI in the United Arab Emirates – is all about the future. There has been head-spinning expansion from a small postgraduate institution four years ago to more than 560 postgraduate students and researchers today, and this August the first undergraduates arrive.
“There’s a lot going on,” admits Professor Timothy Baldwin, an internationally respected AI expert who has worked at top universities around the world. He is provost of MBZUAI, or Mohamed bin Zayed University of Artificial Intelligence, named after the UAE president. An understatement, perhaps.
The university has grown very fast, despite a tight AI focus. The first postgraduate students were admitted in January 2021. Two years ago there were 52 faculty and 202 students. During the 2024-25 academic year, this grew to more than 84 faculty and 365 masters and PhD students as well as more than 200 postdocs and researchers. There are currently seven masters and six PhD programmes.
All students are fully funded. The first undergraduate students will swell the numbers further. Baldwin reckons that at this rate, in a few years MBZUAI will need physical expansion too.
Already the university is ranked highly in the global Computer Science or CS Rankings. For 2024 to 2025 it is number 14 globally in the two specialist areas of computer vision and natural language processing and 29 in machine learning. In AI, the top of the CS Rankings is dominated by Chinese universities.
In the past 18 months, the university has launched multiple initiatives. They include five new departments and the MBZUAI Incubation and Entrepreneurship Centre that supports AI-driven start-ups and has produced 15 so far from faculty and students.
An Institute of Foundation Models was launched in 2023 and supports AI development in the Global South. The university has built major open-source models like K2 and JAIS – an Arabic-centric large language model (LLM) – as well as models for under-represented languages. It has engineering offices in Paris and (soon) the Bay Area in California.
Baldwin spoke to University World News about the “incredible luxury” of having a clean slate when creating a university from scratch. Research and curriculum freedom, generous government funding, top researchers and no grumpy faculty resisting change.
I first interviewed him two years ago, when MBZUAI had been going for two years, and was keen to see how this unusual AI university has been progressing.
Turning internationalisation on its head
MBZUAI and some other new universities in the Middle East are turning higher education internationalisation on its head. They are mostly situated in the politically stable countries of the UAE, Qatar and Saudi Arabia, where governments are investing heavily in universities and research.
With well-funded programmes and research and competitive salaries, they are attracting growing numbers of international students and staff.
Currently, some 80% of MBZUAI’s students are international and about 20% are UAE nationals – usually the international-to-local proportions are the other way around. A similar split is expected at the undergraduate level, although the proportion of UAE students might be slightly higher. Most faculty are also international.
“The undergraduate programme was only announced in March, and in the ‘exceptional talent’ track there are students from countries ranging from Kazakhstan to Brazil, Serbia to China, and Indonesia to America,” said Baldwin, formerly of the University of Melbourne in Australia.
“They really do come from around the world and from all walks of life, which is something we’re emphasising. Our students should not just be classic science wunderkinds,” he noted. The university is also looking for young people with drive and passion, he said, who excel in diverse fields, such as athletes or musicians, and from different cultural backgrounds.
Interestingly, more than 30% of students are female – very unusual for an AI institution.
Are there concerns about having enough local students? Baldwin said there will be around 100 students in the first undergraduate intake. Next year, there will be around 300 students. “So, from a relatively small Emirati population, of around 1.3 million, the absolute numbers become significant very quickly,” he stated.
MBZUAI and the UAE national AI strategy
MBZUAI is an essential part of the UAE National Strategy for Artificial Intelligence 2031, which places AI development at the centre of economic strategy. It aims for the UAE to be a global AI leader by 2031.
Considerable progress has been made. Stanford University’s Global AI Vibrancy Tool ranks the UAE as fifth in the world after the United States, China, the United Kingdom and India. “It’s a fantastic achievement for the country. And we can see the university’s fingerprints all over that,” said Baldwin.
MBZUAI has helped to attract foreign direct investment. IBM set up a collaboration on campus earlier this year, and two major Microsoft centres have been announced in partnership with G42, a local AI company. “We are here, and people want to work with us,” Baldwin commented.
The MBZUAI Incubation and Entrepreneurship Centre supports AI innovation, and like many other researchers, Baldwin is involved in a start-up. LibrAI helps organisations to implement safer, more reliable and better generative AI systems. It does a lot of work in aligning LLMs with local cultures and values and is expanding into agentic AI.
MBZUAI President and University Professor Eric Xing – a world-renowned computer scientist and expert in machine learning, who has taught at top American universities – co-founded the GenBio AI start-up for biological foundation models and research which aims to build the world’s first AI-driven digital organism, AIDO.
The university also does a lot of work in AI training. There is an executive programme as well as bespoke offerings for companies and government entities. “We’ve trained many thousands of people through these programmes,” Baldwin said.
Expanding into undergraduate training
Postgraduate students have the skills to work in a research environment, but the main demand for AI skills is at the lower level, according to Baldwin. The aim of the undergraduate programme is to provide top-flight students with technical skills and business smarts as well as understanding of the world and of the complexities of data and AI problems.
“We were always going to go undergraduate. It was just a question of exactly when we flipped the switch. The reason for doing graduate first was about quality. It’s hard to develop a truly world-class, stand-alone undergraduate programme. It’s much easier to establish the graduate school first and then build down from there, off strengths that you have,” he noted.
In the world of AI, it is unusual to find an academic programme that is not retrofitted to a computer science undergraduate major or degree. MBZUAI asked industry what it needed, and the feedback was that graduates even from top universities around the world are not job ready.
One strategy to address this problem could be to align the core curriculum with what is happening in industry. Another could be to draw on faculty with industry experience. “This is the case for a lot of faculty in the AI space nowadays,” Baldwin said.
Undergraduate students will do meaningful AI work right from the start. They will enter one of two tracks: a BSc in AI – Engineering or a BSc in AI – Business. The engineering track focuses on AI model development and deployment across the breadth of AI in different sectors, while the business track centres on AI business integration and entrepreneurship.
“Cross-disciplinary training and hands-on learning will be the cornerstone of the programme. Students will embrace the humanities, business and a liberal arts curriculum, as they work well outside the bounds of classic computer science and STEM,” the university said when announcing the undergraduate programme in March.
“Furthermore, they will engage in co-operative industry placements, internships, mentorships, and partnerships with top players in industry and AI research.”
MBZUAI has strong connections with industry. It draws on people from industry to teach what is happening with AI at the coalface. And there is a strong focus on experiential learning and on industry placement. Undergraduates will spend one out of four years in industry or AI research contexts, “with a lot of mentoring to make sure they’re learning along the way”.
Staying ahead of the AI curve
Given the rapid evolution of AI, MBZUAI’s links with industry are important for keeping up with the field. Of course, universities around the world are driving aspects of AI development, but there is even more AI research happening in the private sector.
Being a postgraduate-strong university means there are faculty who are innovating at the edge of the field, so academic content is going to be up-to-date. But universities can’t hope to compete directly with research from the likes of Google DeepMind: its team that built the Gemini multimodal LLM has 5,000 people.
“Having said that, we’re one of a very small number of universities in the world that are training commercially relevant major LLM models and foundation models more generally,” he noted.
MBZUAI has been expanding and diversifying its academic programme. In 2024 it started five new departments – in computational biology, computer science, human-computer interaction, robotics, and statistics and data science – and another five will be launched this year.
This represents a shift outside the core AI that includes computer science, natural language processing, computer vision, machine learning and robotics. There will be three broad areas: life sciences, decision science, and the existing AI-focused provision.
The shift to the life sciences led to a School of Digital Public Health. New departments include operations research, precision medicine and systems epidemiology, with faculty spanning both core life sciences and AI expertise.
“The idea is that in co-locating life scientists with the core AI departments, magic will happen. There is so much potential in terms of applications of AI to the life sciences, as we saw last year with the Nobel Prize in biology. That’s the accelerating effect of AI in research,” Baldwin said.
The School of Decision Science will focus on behavioural economics, operations research and policy using AI-driven digital simulations. This year, MBZUAI is also introducing an engineering-focused Master in Applied AI, targeting non-AI graduates seeking real-world AI application skills.
The university has been involved in building foundation models – large and versatile AI models trained on massive amounts of data that may be adapted to many tasks.
Foundation models for biology are being led by GenBio AI. Other achievements are an Arab-centric model called JAIS and a major English language LLM called K2, under the open-source initiative LLM 360. It is aimed at the Global South, to help democratise access to AI development. Models are also underway for under-represented languages in Asia and Africa.
MBZUAI has opened labs in Paris and in the Bay Area, to support global collaboration and attract top talent. “We’re hiring as we speak,” said Baldwin. In five years, MBZUAI expects to have around 400 faculty, 1,000 postdocs and engineers, 1,000 PhD students and 1,000 undergraduates.