Estimated read time: 4-5
minutes

SALT LAKE CITY — Rather than tiptoe around the inevitability of artificial intelligence’s creep into society, the University of Utah has launched a $100 million research initiative that will dig into ways AI can be used responsibly to tackle societal issues.

“AI is something that everybody is talking about,” said Manish Parashar, director of the Scientific Computing and Imaging Institute at the U. “It’s probably the most transformational technology of our time with tremendous potential.”

The Responsible AI Initiative will be led by the institute and will look to advance AI in a way that achieves “societal good” while also protecting privacy, civil rights and liberties and promoting principles of accountability, transparency and equity, the university said.

Before his new role at the U., Parashar was the former office director of the National Science Foundation’s Office of Advanced Cyberinfrastructure and co-chairman of the National Artificial Intelligence Research Resource task force and led the development of the national strategic plan for the Future Advanced Computing Ecosystem as co-chairman of a subcommittee of the National Science and Technology Council.

He said we’re already seeing AI being applied to important problems ranging from the generation of new antibiotics to addressing natural disasters with a seemingly endless stream of other applications and possibilities.

Still, there’s a lot to learn about using AI, especially in a responsible way. We know it can lead to bias and perpetuate societal stereotypes and inequities — something Parashar and his institute is keenly aware of.

So how do we leverage AI in a responsible way to be able to impact these grand challenges, which are so important to this region, to Salt Lake City, to Utah, to the Intermountain West?

–Manish Parashar, Scientific Computing and Imaging Institute director

“When we think about, ‘How do we advance AI?’ we have to take these potential downsides of AI into consideration and build in the right guardrails so that we are advancing AI in a responsible way. This is really what this initiative is trying to do,” Parashar said.

The initiative will also have a focus on harnessing AI to help with Utah and region-specific issues like public services, health care, sustainability, water resources and building the right workforce for the jobs of the 21st century.

“So how do we leverage AI in a responsible way to be able to impact these grand challenges, which are so important to this region, to Salt Lake City, to Utah, to the Intermountain West?” Parashar said.

The initial phase will see deep technological expertise, advanced cyberinfrastructure and disciplinary expertise across the university brought together through the initiative, positioning the U. as a national leader in translational AI. Parashar said this will include researchers in the fields of sciences, engineering, medicine, humanities, social and behavioral sciences, policy and social work and more, working together.

Initial funding for the initiative will raise and repurpose funds from three non-tuition sources: returned overhead, investment income and philanthropy and university leaders expect the initiative to eventually generate additional, focused future funding. Because of the energizing nature of AI research, future gifts to support the project and additional areas of interest are possible, the university said.

“You have to do this in a way that protects (people’s) privacy, protects civil rights, civil liberties so that you’re guarding against negative impacts here,” Parashar said. “Looking at problems that are interesting and important to this area and bring the expertise we have … to be able to address them in a responsible way. That’s what will allow us to differentiate ourselves from everybody else that is also looking at AI and allow us to become a leader.”

But what exactly does protecting people’s privacy and rights look like in practice?

Parashar said doing so means vetting data sets to see if they have bias and having oversight mechanisms and protocols in place to flag specific functions or applications of AI that don’t look right so that corrective action can be taken, among a host of other things. Essentially, build solutions to these issues into the technology of AI.

“As part of the project, the U. will establish an internal governance council and external advisory board of national and global AI leaders to provide advice and guidance, as well as expand its faculty by hiring clusters of experts focused on grand challenges. The investment will also include enhancing faculty support structures and building a cutting-edge cyberinfrastructure that will advance AI capabilities globally,” according to a release from the University of Utah.

“By creating a widely accessible advanced cyberinfrastructure that ties computational resources, data, testbeds, algorithms, software, services, networks and user training and expertise, the initiative will create new opportunities for progress across all fields and disciplines. Increased access will create opportunities for ethical AI guardrails, including AI auditing, testing and evaluation, bias mitigation and safety,” the release said.

Like other U. leaders, Parashar is extremely optimistic about the direction this initiative can steer AI in.

“I think we can look at a future where we can really have the right technologies that are helping us better navigate our lives, better manage our own personal health (and) create solutions for many of these global problems you’re seeing that are so urgent at our time,” Parashar said. “I see a lot of that coming together by building on this approach that they’ve laid out in this initiative.”

×

Most recent Artificial Intelligence stories