LAS launches new interdisciplinary initiative on trustworthy AI

Artificial intelligence is everywhere these days whether we recognize it or not – literally. How many of us have been duped by thinking a video or photo is real, only to find out that it’s AI-generated? Or we ask Google a question and get an AI response and wonder, “Should I trust that answer or keep looking?”

Faculty in Iowa State University’s College of Liberal Arts and Sciences (LAS) are at the forefront of the AI technical revolution, developing computer solutions, quantifying the statistical certainty of AI outputs, and exploring ways for AI to enable new inventions and innovations.

Importantly, LAS scholars and researchers are also studying the ethical implications and trustworthiness of AI, which helped to inspire a new initiative called Trusted Innovation in AI.

Applying unique LAS strengths to AI

The Trusted Innovation in AI initiative primarily focuses on research that builds on LAS’ strengths in trustworthy, ethical, and interdisciplinary AI. The college is also exploring broader AI efforts, including integrating AI into curricula and establishing policies on how to implement AI in the classroom.

“Our faculty in the College of Liberal Arts and Sciences are uniquely positioned to lead research on trustworthiness in artificial intelligence because they build from a perspective that the challenge of generative AI is not solely technical — it is fundamentally human,” said Benjamin Withers, dean of LAS. “Our strengths across the humanities, social sciences, mathematics, and natural sciences allow us to examine AI through the lenses of ethics, communication, policy, human behavior, and scientific rigor simultaneously. By bringing these disciplines together, we can help ensure that AI systems are not only innovative, but transparent, accountable, and worthy of public trust.”

Within the Iowa State University AI ecosystem, LAS scholars contribute expertise in developing computing solutions, quantifying the statistical certainty of AI outputs, exploring how AI enables new discoveries and innovations, and studying the ethical and societal implications of AI.

For LAS, the goal is to not only create new AI applications, but to ensure those applications are trustworthy and ethical.

A need for trust and ethics in AI

A common complaint about AI is something called hallucinations, information generated by AI models that sounds true and feasible but is actually false and incorrect.

“A lot of people working in the foundational AI portion are interested in knowing when the answer is correct, like our scholars in statistics, the Center for Survey Statistics and Methodology, and computer science – that is what they are doing. They want to know if what we get out of AI is correct,” said Gustavo MacIntosh, LAS associate dean for research and graduate education.

Another component of trust in AI is determining how models come up with answers to questions.

“Another aspect is explainability,” said Geetu Tuteja, LAS interim assistant dean for strategic research initiatives. “People need to understand how an AI system arrives at its answer and what data or reasoning informed those results.”

Closely aligned with trustworthiness in AI is ethics. Innovative advancements in the technical aspects of AI are exciting, but when does it go too far? For example, the news is filled with instances of artists’ likenesses being used without permission. There’s also the reality of AI’s impact on the environment: How will growth in AI affect sustainability concerns?

“LAS is uniquely positioned to approach AI from multiple perspectives, with researchers studying everything from the technical foundations and applications of AI to its ethical, political, and human dimensions,” Tuteja said. “That breadth allows us to examine not only what AI can do, but also how it should be used responsibly.”

A firm foundation

Over the past few years, LAS has developed AI projects focused on trust and ethics, while also creating new AI development opportunities for students. This work, detailed below, is the foundation for future projects related to the Trusted Innovation in AI initiative:

• Researchers at the Center for Survey Statistics and Methodology are developing AI-assisted methods to support the National Resources Inventory (NRI), a nationwide program that monitors land use and natural resource conditions across the United States. The project uses advanced AI and computer-vision techniques to help detect changes in land cover and land use categories from high-resolution imagery, while also quantifying uncertainty in the AI predictions.
• A new minor in applied AI teaches undergraduate students how to use AI to address real-world challenges. Students also explore societal implications, ethical considerations, and potential challenges associated with AI.
• A Master of Science in AI offered through the Department of Computer Science equips graduate students with a deeper understanding of AI and machine learning. This degree helps students apply AI to important global challenges.
• The Dependable Data-Driven Discovery (D4) program is an interdisciplinary research and training initiative focused on ensuring rigorous quality assurance for data-driven and AI discoveries originating from biological and other data sources. Graduate students may apply for traineeships and undergraduates may apply for assistantships to learn about data science foundations, methodology, and tools for creating trustworthy data-driven discoveries.

Future funding

New projects related to the Trusted Innovation in AI initiative are in the works and will launch once funding becomes available. New financial support will also help fuel collaborative workshops and future ideas.

Currently, most of the activities for the Trusted Innovation in AI initiative are funded by private donors in the form of named professorships. Current AI-related professorships held by LAS faculty include:

• Peng Liu, professor of statistics, Laurence H. Baker Chair in Biological Sciences
• Carole Chappelle, professor of English, Distinguished Professor and LAS Dean’s Professor
• Wallapak Tavanapong, professor of computer science, Kingland Systems Professor in Data Analytics and Cognitive Machine Learning
• Myra Cohen, professor of computer science, Lanh and Oanh Nguyen Chair in Software Engineering
• Zhengyuan Zhu, professor of statistics, Kingland Systems Fellow in Data Analytics and Cognitive Machine Learning, director of the Center for Survey Statistics and Methodology, co-director of the ISU Federal Statistics Research Data Center

Tuteja is spearheading this collaborative movement and recently organized a luncheon to connect LAS faculty who are interested in developing AI partnerships. LAS faculty who want to join future events or find partners for new AI projects should contact Tuteja at geetu@iastate.edu.

“We see this as a seed to produce something bigger,” MacIntosh said. “We are bringing them together, helping them connect, and providing some support for leadership.”