Shishir Shah’s new office was bare, aside from a few textbooks on the bookshelf and a balloon sign that read “Welcome Dr. Shah,” a reminder of his new position as OU’s first university-wide chief artificial intelligence officer.
“Truthfully speaking, I didn’t have much time to visit Norman before I joined (OU),” Shah said. “We’re still in the process of finding a place to stay. Most of our evenings are spent looking for a home.”
In August, Shah was announced as OU’s new chief artificial intelligence officer and director of the School of Computer Science.
Before coming to OU, Shah spent over 20 years in the department of computer science at the University of Houston as a professor and in various leadership positions.
Shah also directed research at the University of Houston’s Quantitative Imaging Laboratory, focusing on an AI field called “computer vision,” where machines are able to process and make decisions based on videos and images.
“A significant portion of what we perceive is what we see. … That kind of sensing is critical in us doing a variety of cognitive-level tasks,” Shah said. “The question is, ‘Can we figure it out and allow computers to mimic that process?’”
Shah attended the University of Texas at Austin, where he earned his bachelor’s degree in mechanical engineering in 1994, his master’s degree in 1995 and his doctorate in electrical and computer engineering in 1998.
As an undergraduate student, he designed robotic arms, actuators and feedback control systems.
“In order to do feedback-driven manipulation, you needed sensing. Then I got exposed to using cameras as a way to do perception and sensing of the environment,” Shah said.
After joking that he originally did it for the money, Shah said his academic career was mostly influenced by working with faculty on research as an undergraduate student.
“My first involvement in research with the faculty members was mostly driven based on the fact that I could get a monthly stipend for it,” Shah said. “But in hindsight, I would argue that it helped me connect a lot of things that I was learning in my courses to something that was translatable into solving problems (and) gaining hands-on experience.”
Future of AI at OU
In 2020, OU launched “Lead On, University,” an eight-year strategic plan centered around five pillars to represent significant goals set by the university. As part of the plan, which was refreshed in March, the university aims to prepare students to navigate artificial intelligence effectively after graduation, according to a mass email announcing Shah’s appointment in August.
“By advancing a comprehensive, university-wide strategy to lead in AI innovation, research, education, and implementation, we are reimagining curriculum and academic opportunities to integrate AI across the student experience,” the email read. “This ensures our graduates enter the workforce not simply as literate AI-users, but as responsible, informed ones who know how to leverage this technology for maximum impact.”
Shah said OU is one of the few universities strategically implementing AI in its plan for the future.
“As universities grapple with the presence of AI within our environments, universities need to think about, ‘How does this affect our teaching mission?’” Shah said. “Having an ability to strategically think about how we integrate AI and how we make it available so that we can support what universities need to do.”
About a quarter of the university’s research is rooted in AI, according to the mass email. Shah said the university’s AI research ranges from developing AI systems to integrating AI into ongoing research.
Shah said his goal is to provide additional resources to faculty members to continue their research and share knowledge across all OU campuses.
“If we can provide a way so that it’s more easily shareable information, those experiences can be translated and help someone jumpstart and accelerate their research efforts,” Shah said.
Shah said responsible AI use includes understanding the capabilities of AI and holding a foundation of ethics.
“Sometimes it becomes difficult to understand whether we are using (AI) to complement or augment what we are doing, or are being driven by what AI is suggesting,” Shah said.
Shah said that AI will ideally be integrated into degree programs at different scales to increase exposure and engagement with AI.
Pillar five of the strategic plan outlines tactics to “leverage OU’s AI-related expertise across disciplines and in research centers and institutes.” The strategy also aims to position Oklahoma to be a leader in AI fields.
“Certainly for academia, it’s changing the way we are learning (and) the way we are able to access information, … but there are aspects of this that are bringing efficiency to what we used to do,” Shah said.
Shah said that students will be taught about the environmental impact of generative AI.
AI is largely powered by data centers that store data and deploy information, and as AI increases in popularity, the demand for data centers increases. These centers generate heat and they rely on fresh water to stay cool.
Larger centers can consume up to 5 million gallons of water per day, according to the Environmental and Energy Study Institute.
“Computer resources have an environmental impact, just like almost every other thing does,” Shah said. “We’re consuming a lot of energy in order to see something happen at our fingertips in real time.”
Shah said that an element of responsible AI use includes fact-checking information that is generated by AI.
“We want to be able to build an understanding of how to fact-check and how to understand what’s right and what’s not right,” Shah said. “That part is a learning experience because it’s new to us and, in some cases, it may seem redundant.”
Shah compared the use of generative AI to the invention of calculators, stating that it is still important to know how to do math if there is a calculation error, similar to how he sees responsible AI use.
“There’s a balance to be found somewhere along the way,” Shah said. “We have already found that balance with calculators. In time, we’ll find our balance with these large language models.”
This story was edited by Ana Barboza, Anusha Fatherpure, Natalie Armour and Thomas Pablo. Kennedy Johnson, Avery Avery and Mary Ann Livingood copy edited this story.
