Students are not the only ones struggling to learn about artificial intelligence. Universities face big challenges in preparing the next generation for careers that use AI, robotics and data science. They face a long list of hurdles, from attracting and retaining the right faculty and developing curricula to providing the AI computing infrastructure for students to perform research and learn new skills.
In the US, the number of students trying to learn computer science far exceeds the supply of professors. So how can universities prepare their campus, faculty and students for a world that will increasingly rely on computer science and artificial intelligence?
The University of Florida has pioneered a replicable framework for future public-private cooperation and a network of AI-driven institutions. The first university in the country to introduce AI curricula across all disciplines, it has hired more than 100 AI faculty members to deliver this.
To achieve this, the university worked with NVIDIA to develop necessary computing hardware, software, training and services to become an “AI university”. The approach has been highlighted by the Center for Data Innovation as an example of a successful industry-university partnership, resulting in a supercomputer that will be shared across 12 public universities in Florida, including Florida A&M – the largest HBCU in the country – and Miami-Dade College, and unprecedented access to AI training and tools for under-represented communities.
- Spotlight collection: AI and the university
- Digital first: how to use AI to enhance customer service in universities
- THE podcast: is AI in higher education worth the hype?
The blueprint
The success of this partnership relied on a precise, detailed road map – one that can be used by colleges and universities across the country. There are takeaways for other universities to learn from when building AI capacity.
First, both parties need to clearly articulate realistic goals, based on the current institutional environment, skills and systems. Setting these goals according to the strengths and missions of the two partners should be a priority.
For instance, when the Florida team began setting goals, they had a decade of experience maintaining a supercomputing environment and the physical space to house a new AI supercomputer. Other universities might not have experience acquiring or managing high-performance computing systems. They can still bolster their AI capabilities through smaller, mid-range or cloud-based systems with a focus on improving specific systems or processes such as careers advice or student support.
Second, focus on increasing AI computing capacity at the university. If AI computing resources are not available, there is nowhere for AI innovation and development to take place. Currently, few universities have invested in this infrastructure because of lack of funding and resources. Partnering with industry and government can enable a university to overcome this barrier.
Third, all parties should focus on how to increase buy-in and interest across the university. This can be challenging because most colleges within an institution have a significant amount of autonomy and their own budgets. A university-wide project might reallocate funds, so early buy-in means inviting everyone to be a part of the process.
The University of Florida asked each department to build its own core AI curriculum, including components specific to its needs. Spreading opportunities throughout disciplines using an AI across-the-curriculum approach engaged every college in the process.
Fourth, evaluate and measure the success of the programme, once it is up and running, on an ongoing basis. The benefits might be difficult to quantify initially, but benchmarks and measurable impact are vital for future partners, prospective faculty or researchers to understand the relevance and value of the programme.
Florida is in only its second year of the initiative and is working to compile its success metrics, but it has set out a few broad goals, including capturing the number of students who go through all three AI classes that are available to everyone, and assessing how the AI classes are helping with student recruitment. It is also looking at ways to measure how the school’s new AI opportunities are increasing the quality of education and attracting new students and faculty.
Since deploying its supercomputer and beginning to offer AI curricula in 2020, the university has seen several positive developments, including:
- Faculty recruitment: 130 AI-focused faculty from diverse backgrounds have been recruited to teach AI techniques throughout the university, including the College of Arts, the College of Medicine and the Institute of Food and Agricultural Sciences
- Increasing access to underserved groups: The supercomputer is available for use to 12 public universities in Florida
- Building an AI-fluent community: More than 1,000 faculty, students and researchers have received certified, instructor-led AI training.
Time for action
AI may be the most transformative technology of our lifetimes, with the potential to generate huge economic value and address serious challenges to society, such as improving healthcare and mitigating climate change. The world’s leading organisations, including Amazon, Microsoft, Tesla and many others, are racing to deploy more effective products and services that are powered by AI, meaning the demand for computer science researchers, graduates and faculty is unprecedented. Universities need to extend their AI expertise to help prepare students for these rewarding careers.
Cheryl Martin is director of higher education research at NVIDIA.
If you found this interesting and want advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the THE Campus newsletter.
comment