From idea to implementation: How ASU closes the AI gap

Editor's note: This expert Q&A is part of our “AI is everywhere ... now what?” special project exploring the potential (and potential pitfalls) of artificial intelligence in our lives. Explore more topics and takes on the project page.

Program manager Paul Alvarado of the AI Acceleration team at Arizona State University takes a practical approach to his work, helping to bridge the gap between feasible ideas and their implementation. This work extends to artificial intelligence, for which accurate data is crucial.

"It’s common for people to have great ideas, but implementing them is another matter," Alvarado said. "This is where we get into the details, including figuring out if we have the right data and how the data is being updated."

Alvarado has been with ASU Enterprise Technology for five years. Just this year, he joined the AI Acceleration team, and his data expertise is shaping the university’s approach to AI.

Here, he shares his approach to narrowing the AI gap in higher education. 

Question: What is the focus of your work to advance AI at ASU? 

Answer: I’ve been specifically assisting with the AI Innovation Challenge, which is conducted in collaboration with the Office of the University Provost and our partners at OpenAI.

We initiated this challenge to explore AI and its impact in the classroom, workforce and research. This challenge allows ASU to have an impact-driven approach to using the ChatGPT Enterprise licenses as people across ASU are all thinking about how this technology can impact the work they do, and they will tell us about how it went upon completion of the challenge. This provides us the opportunity to review and analyze their activities and experiences, including how they started and how things have progressed with the technology.  

My latest project is supporting the AI Innovation Challenge, a collaboration with the Office of the University Provost and our partners at OpenAI, which allows students, staff and faculty to use the ChatGPT Enterprise licenses.

Once people across the university find opportunities to implement the tool, they can then share the impact of their progress and results.

This provides us the opportunity to review and analyze their activities and experiences, including how they started and how things have progressed with the technology.

Q: What role does data play in this work?

A: Data is fundamental to everything we do at ASU, and it is equally — if not more — important in AI.

To make the AI Innovation Challenge successful, ASU has established feedback touch points with proposal winners to understand how to best support colleagues across ASU. We have found that folks want to be able to enter in data from public ASU websites, knowledge bases, policy documents, etc., as opposed to loading a single document that becomes stale over time. Through this we have learned that establishing data connections is important to stakeholders, so Enterprise Technology must establish processes with vendor AI products to have these connections reviewed from a security standpoint. We want to enable innovation but do so responsibly by making sure we have analyzed how data is passed to and within a vendor solution.

Outside of the AI Innovation Challenge, we have been incorporating data from internal knowledge bases (ServiceNow and Guru), ASU public websites, and documents (Google docs and Microsoft Word docs). What we have seen is that GenAIGenerative artificial intelligence can leverage these sources of knowledge to answer questions, but some obstacles must be navigated. 

Examples of obstacles are making sure the sources of truth are maintained/updated, identifying and addressing conflicts between sources of truth and creating articles or information if the answer to a particular question doesn’t exist. We also see that important information can be placed in images, presentations, tables, etc., that don’t always easily translate to AI-ready content, so best practices for storing knowledge also have to be revisited.

Q: How can AI tools redefine the role of educators, particularly in terms of taking over routine tasks and enhancing teaching resources?

A: One way we can help instructors reduce routine tasks in the classroom is with Syllabot, a unique AI experience built using MyAI Builder. It is an AI-powered chatbot that helps students navigate the syllabus for a specific course.

In addition to working directly with faculty, students can input questions into Syllabot to understand the course schedule, assignment breakdown and more, as outlined in the syllabus. Our team collaborated on this project with Danielle McNamara, a professor of psychology and executive director of the Learning Engineering Institute.

The Syllabot project helps narrow the scope of our work by focusing on specific data points. Using Syllabot, we can identify its capabilities and limitations. Then, we can partner with instructors to determine improvements. If Syllabot cannot easily answer questions based on the course syllabus, the syllabus may need reformatting or content restructuring. To optimize tools like Syllabot, we must experiment with different approaches and learn from the outcomes.

In July, we’re going to test Syllabot in an accounting class taught by Euvin Naidoo of ASU’s Thunderbird School of Global Management and plan to test it further in another class in the fall. With more feedback and adjustments, including student interactions with Syllabot, we’ll be able to refine and expand its use. However, ensuring we have the right feedback and data is crucial, as always.

Q: How is ASU advancing the successful development of AI use in higher education?

A: We help advance AI in higher education through initiatives like the AI Innovation Challenge launched earlier this year. We've been able to work with OpenAI to provide them with feedback on our early experiences. This type of collaboration allows us to inform industry leaders about their products. The value of early and steady feedback in partnerships like this is part of ASU's DNA when it comes to advancing innovation.

With the OpenAI partnership, ASU aims to scale with Principled Innovation to ensure we’re doing the right thing for society and the future. We want to scale responsibly. 

For example, when we partnered with OpenAI for the ChatGPT Enterprise and later helped shape OpenAI’s ChatGPT Edu, one of the biggest challenges was managing the scaling at our level — such as provisioning and deprovisioning user access at scale. With students and faculty members frequently entering and leaving our system, manually managing permissions is not feasible. We worked with OpenAI to automate this process effectively, ensuring our system integrates seamlessly with theirs. This not only helps us but also benefits countless other institutions of higher learning.

Q: What opportunities related to AI in higher education most excite you?

A: What we’re doing here at ASU is extremely special. With many AI vendors out there, it’s great that we can form partnerships with innovative entities like OpenAI. We focus heavily on the outcomes of technology tools, and often the evaluation becomes an afterthought.

At ASU, we build in steps to analyze our processes continuously. We have access to all our own data and don’t have to wait to get it back from the vendors we work with. It’s an exciting opportunity — having real-time access to our data allows us to move toward evaluation and improvements much more quickly.

AI is everywhere ... now what?

Artificial intelligence isn't just handy for creating images like the above — it has implications in an increasingly broad range of fields, from health to education to saving the planet.

Explore the ways in which ASU professors are thinking about and using AI in their research on our special project page.