Guiding Responsible AI Innovation at the Nexus of Research and Real-World Impact
With an academic and professional journey that spans across Rutgers University, the National Science Foundation, the White House Office of Science Technology and Policy (OSTP), and now the University of Utah, Manish Parashar, Ph.D., has followed a path driven by a desire to make a meaningful impact and has sought roles where research, policy, and infrastructure converge to drive scientific advancement and societal progress. “I’m an engineer at heart who enjoys research in computing, but especially the ability to impact science and society through research,” shares Parashar, Inaugural Chief Artificial Intelligence Officer, Director of the Scientific Computing and Imaging (SCI) Institute, Chair in Computational Science and Engineering, and Presidential Professor in the Kahlert School of Computing, University of Utah. “This vision has driven my career trajectory, and I've transitioned towards opportunities that allow me to have that impact through my research, but increasingly through shaping policy, building cyberinfrastructure, and developing strategies that influence both science and society.”
Advancing Responsible AI
As the inaugural Chief AI Officer at the SCI Institute and a Presidential Professor, Parashar is leading the strategic alignment of Utah’s efforts to advance responsible AI across research, education, and policy. “AI is one of the most revolutionary technologies of our time,” says Parashar. “This technology is evolving quickly and enabling us to accomplish things on an unprecedented scale. If we apply it strategically and collaboratively, it can help us solve pressing challenges here in Utah. The goal of the Responsible AI Initiative is to bring together multidisciplinary, sociotechnical teams that can holistically address critical challenges for the state, the region, and the country and align our strengths in technology, policy, law, and the social dimensions of AI.”
“This technology is evolving quickly and enabling us to accomplish things on an unprecedented scale. If we apply it strategically and collaboratively, it can help us solve pressing challenges here in Utah. The goal of the Responsible AI Initiative is to bring together multidisciplinary, sociotechnical teams that can holistically address critical challenges for the state, the region, and the country and align our strengths in technology, policy, law, and the social dimensions of AI.”
Part of the SCI Institute is the One-U Responsible Artificial Intelligence Initiative (One-U RAI), which aims to advance AI to achieve societal good, all while protecting privacy, civil rights, and civil liberties, and promoting fairness, accountability, and transparency. “The Initiative has four pillars, with the first one being building expertise,” explains Parashar. “We want to bring together faculty, visitors, postdocs, and researchers to form teams that align our existing strengths at the University of Utah while also bringing in new expertise. The second is infrastructure. AI depends on computing and data, so we’re focused on providing the right capabilities to help these teams innovate and succeed. The third pillar is creating structures that facilitate multidisciplinary collaboration and structures that act as catalysts for research and innovation. It’s about creating opportunities and environments that support that kind of cross-disciplinary work. And the fourth is engaging with society. To be truly responsible, we not only need to build the technologies and solutions, but also bring the public along and communicate, engage, and make sure they’re part of this effort, in Utah and beyond.”
“The philosophy behind all of this is to build on the unique strengths we have at the University of Utah,” Parashar adds. “We have a history of translational impact, access to unique datasets, and a deep understanding of regional issues. The Responsible AI Initiative is about leveraging those strengths to address the challenges that matter most.”
Pushing Advanced Technology Forward
Building on the four foundational pillars of the Responsible AI Initiative, Parashar says this work aims to drive real-world impact across Utah, in areas like education, environmental health, and water sustainability, and position the state as a model for national AI innovation and collaboration. “In the area of education, we have a significant urban, rural, and frontier divide in Utah, with big gaps in digital access and capabilities. So how do we use AI not only to teach about this technology, but to teach with AI? How do we use it to amplify and accelerate the ability of educators to deliver more effective instruction, whether in high schools, universities, or vocational programs? At the same time, we need to prepare students for the future of work. AI will touch almost every job, so it’s essential that we provide people with the skills to be competitive in the 21st-century economy.”
“We've long envisioned infrastructure as a utility, something you don't have to think about, and we're getting closer to that. But the next step is to start thinking about AI as infrastructure, not just a tool we use, but something embedded in everything we do. That’s where the future lies. We are in an incredibly exciting time, and while we are facing serious national and global challenges, we’re also gaining the tools to begin addressing them in meaningful ways. AI is becoming a powerful part of the toolkit we can use to take on these complex problems. And if we use it responsibly, I truly believe we have a chance to achieve the kinds of societal benefits we’ve always dreamed of, and meaningfully improve lives, communities, and our shared future.”
AI is also being applied to address some of Utah’s most pressing environmental challenges. “Air quality is a critical issue here, especially given our geography and the effects of inversion,” says Parashar. “We’ve collected valuable datasets over the years, and we’re using AI to better predict, manage, and communicate the impacts of poor air quality, including how it correlates with health outcomes. Water sustainability is another key area. We have extensive expertise in subsurface hydrology and years of data, and AI gives us a powerful tool to translate that understanding into real-world solutions with immediate impact.”
“Our approach is unique in that we’re using AI to solve real problems, and as we do that, we’re pushing the technology forward,” continues Parashar. “This kind of applied, translational work not only helps people now, but also contributes to advancing AI itself. I’ve found it remarkable since joining the University of Utah how easy it is to collaborate, not just across disciplines, but with government and industry. We’ve created the Responsible AI Community Consortium, which brings together representatives from academia, the state, and industry to look at challenges from a truly holistic perspective. Through our partnership with Utah's Office of AI Policy, we can do innovative research within an 'innovation sandbox.' That means if a researcher is testing a new algorithm or technology, they can apply for regulatory mitigation within that controlled environment. It allows us to move quickly and responsibly from research to real-world impact. These kinds of public-private partnerships are crucial for national progress, because it’s not about competing with industry, it’s about collaborating with them to drive innovation, and this model is something that can scale beyond Utah.”
Shaping the Next Generation of AI Leaders
As AI continues to evolve and influence nearly every aspect of society, higher education plays a critical role in shaping how future leaders engage with and advance these technologies. “One of the core responsibilities of the university is to prepare the next generation, not only to use today’s technologies, but to understand, create, and shape the technologies of tomorrow in a responsible way,” shares Parashar. “We often focus on finding the right technical solutions or identifying the right problems to solve, but we don’t spend enough time considering how the technology will actually be used, and the broader sociotechnical implications that come with developing and deploying AI.”
“We must go beyond just putting datasets online. Data accessibility means building tools, infrastructure, and systems that allow everyone to use them. This approach is the only way we ensure AI reflects the broader community. AI is shaped not only by how it’s trained, but by how it’s used. And that use, whether by educators, researchers, or communities, depends on equitable access to data and the ecosystems that support the sharing, management, and utilization of that data. We also need to build community awareness and develop the skills necessary to help future generations use data effectively. Making data more valuable and usable for users is where we can really make a difference as technologists."
“The more important questions often are: Should we be doing this, and what are the ethical considerations?” continues Parashar. “With AI, these questions are becoming increasingly critical. It’s important to me to help students and early-career researchers think through that full spectrum, not just the technical challenge, but who will use the technology, how it will be used, and what the broader impact will be. I’ve been fortunate to collaborate with so many incredible students and researchers; it's truly the most rewarding part of my job.”
Dr. Forough Ghahramani, Assistant Vice President for Research, Innovation, and Sponsored Programs, Edge, admires Parashar’s commitment to mentoring graduate students and postdocs, and his ability to inspire collaboration across disciplines. “I had the honor of working in Dr. Parashar’s group during his time at Rutgers, where I witnessed firsthand his deep expertise, visionary leadership, and lasting influence on the national cyberinfrastructure strategy and the research community. His ability to integrate technical excellence with ethical stewardship positions him as a national leader in AI-driven science and policy, and he has left an indelible mark on the field. Dr. Parashar’s current work at the University of Utah, especially leading the national conversation on Responsible AI, continues to elevate the impact of inclusive, ethical, and future-forward innovation.”
Approaching Problems Holistically
As quantum computing becomes increasingly relevant to both AI and high performance computing (HPC), Parashar notes the importance of actively exploring its potential and how it may shape the future of research. “Quantum computing is a tremendously promising technology, but we’re still trying to figure it out, both at the device level and the computing level. We’ve made impressive advances in recent years, including creating quantum systems that we can experiment with and quantum accelerators on the horizon. There are already hybrid classical-quantum platforms available, and we’re learning more every day. Our approach is to stay informed, understand the potential, and explore how quantum can complement everything else we’re doing in AI and HPC.”
“Personally, I’m very interested in how quantum systems could help accelerate problem-solving, whether it’s data processing, drug discovery, or other areas where there’s a lot of promise,” continues Parashar. “As the technology matures and more stable systems become available, I believe we’ll need to incorporate quantum into our broader research ecosystem. At the University of Utah, we don’t yet have a quantum system, but we’ve been using external platforms and are actively exploring when it would make sense to integrate it into our infrastructure.”
With experience in both national infrastructure and university-led initiatives, Parashar views Research and Education Networks (REN), such as Edge and CENIC, as critical enablers of shared AI resources, collaborative research, and greater connectivity across regional and national infrastructure. “Science is no longer isolated. RENs are the backbone that provide the essential connectivity for data to move from instruments and simulations to end users. But even more importantly, RENs connect people, acting as catalysts for collaboration and enabling researchers to approach problems from a truly end-to-end perspective.”
“By connecting all of the pieces, we’re now able to approach problems more holistically, whether it’s addressing natural phenomena, understanding complex processes, or responding to events,” continues Parashar. “That means bringing together not just people, but also data, infrastructure, and capabilities. We wouldn’t be able to do any of this without the Utah Education and Telehealth Network (UETN), it’s truly the backbone of everything we do here. The same is true in New Jersey with Edge. These networks are not just enablers; they’re catalysts for innovation.”
Drawing on his leadership as co-chair of the National Artificial Intelligence Research Resource (NAIRR) Task Force and his experience guiding national strategic computing efforts during the pandemic, Parashar is focused on positioning Utah’s cyberinfrastructure and AI ecosystem as a vital contributor to a broader national network by linking universities, government, and industry in meaningful collaboration. “We must first determine how to align with national priorities—whether it's economic development, national leadership in AI, or national security—while also representing and building on our unique regional strengths,” explains Parashar. “We want to amplify what makes Utah and the Mountain West distinctive, and at the same time, contribute meaningfully to national goals in science, technology innovation, and security.”
“Everything I’ve worked on, from national strategic computing to NAIRR, has really been about ecosystems,” continues Parashar. “These aren’t centralized systems with a single purpose. They’re federated efforts, built by bringing together the best capabilities from each region to create something greater than the sum of its parts. That’s the vision we’re following here in Utah. It’s not just about building infrastructure, it’s about building connections that scale impact nationally.”
Improving Data Accessibility
Throughout Parashar’s career, he has led numerous efforts to advance collaborative data infrastructure, including the Virtual Data Collaboratory, the Science Data Exchange (sciDX), and the National Data Platform (NDP), and has gained valuable insights into the complexities of data sharing. “There are extensive technical challenges to data sharing, but also significant sociotechnical ones,” explains Parashar. “Data often contains sensitive or protected information, and there’s a natural hesitation to share it. On top of that, the mechanics are difficult; data is heterogeneous, lacks standardization, and it’s hard to sustain both technically and financially. Right now, there’s no cohesive national data strategy. There are pieces of it, especially in areas like AI, but what’s missing is a unified approach to lowering barriers, adding value to data, and making it easier for people to actually use.”
“This is not just about making data open,” continues Parashar. “That’s important, but it’s often not enough. I always go back to an example from my time at Rutgers, where we led infrastructure for the Ocean Observatories Initiative. We had high-definition ocean floor cameras streaming incredible footage in real time. Oceanographers were thrilled, but at the same time, we got emails from smaller universities saying, ‘We can’t download or store this data, can you just load it on a hard drive and FedEx it to us?’ That showed me the real problem: accessibility.”
Parashar says if you want to democratize AI, data accessibility is critical. “We must go beyond just putting datasets online. Data accessibility means building tools, infrastructure, and systems that allow everyone to use them. This approach is the only way we ensure AI reflects the broader community. AI is shaped not only by how it’s trained, but by how it’s used. And that use, whether by educators, researchers, or communities, depends on equitable access to data and the ecosystems that support the sharing, management, and utilization of that data. We also need to build community awareness and develop the skills necessary to help future generations use data effectively. Making data more valuable and usable is where we can really make a difference as technologists."
Parashar says an entirely new workforce is emerging, shaped by the growing demand for human-centered roles in data and computing. “Data and compute will need to be integral to everything the next generation does, and it must be seamless. We've long envisioned infrastructure as a utility, something you don't have to think about, and we're getting closer to that. But the next step is to start thinking about AI as infrastructure, not just a tool we use, but something embedded in everything we do. That’s where the future lies. We are in an incredibly exciting time, and while we are facing serious national and global challenges, we’re also gaining the tools to begin addressing them in meaningful ways. AI is becoming a powerful part of the toolkit we can use to take on these complex problems. And if we use it responsibly, I truly believe we have a chance to achieve the kinds of societal benefits we’ve always dreamed of, and meaningfully improve lives, communities, and our shared future.”
“The more important questions often are: Should we be doing this, and what are the ethical considerations? With AI, these questions are becoming increasingly critical. It’s important to me to help students and early-career researchers think through that full spectrum, not just the technical challenge, but who will use the technology, how it will be used, and what the broader impact will be. I’ve been fortunate to collaborate with so many incredible students and researchers; it's truly the most rewarding part of my job.”
- Manish Parashar, Ph.D.
