Two students smiling together at a table in a full room.

Hacking for humanity: ASU students create, explore in generative AI

ASU students from the humanities and computer sciences team up to solve world problems using artificial intelligence (AI).

Krista Hinz

October 19, 2023

On any given Friday afternoon, the campus at Arizona State University (ASU) is slowly winding down for the weekend. However, last week the Creativity Commons on ASU’s Tempe campus told a different story.

In the atrium gathered 30 students, along with nearly a dozen faculty and staff, for a two-day hackathon that centered on discovering the possibilities and testing the boundaries of generative artificial intelligence (AI).

The event, called Prompt Engineering Hackathon for Humanities, was hosted by ASU's Enterprise Technology, and invited students to use generative AI tools — like ChatGPT, Bard, and Adobe Firefly — to get a better understanding of the capabilities and limitations of these systems.

The objective was for students to build storified personas – through text, imagery and narrative — to test the limits of generative AI tools and gain a deeper understanding of how humans and machines collaborate.

Plugging into AI prompt engineering with faculty experts

Time spent at the hackathon was split between education and exploration. Faculty and staff mentors shared best practices for prompt engineering — which is the skill of writing clear instructions that can be interpreted by the generative AI tool. With a working knowledge of how to use these tools, students were invited to play, create and problem solve using AI.

Professor Andrew Maynard and Professor Elizabeth Grumbach lead sessions on prompting during Friday’s workshop.

“What's so exciting about the hackathon is that these students are going to get this primer in prompt engineering, but also some exercises on how to engage critically and ethically with these generative AI technologies,” said Grumbach.

With a clearer understanding of how to use AI tools, students self organized into teams to start getting to work. On Friday, ten teams emerged.

For Seher Shah, an English graduate teaching associate, this was her very first hackathon. She commented that the event helped her think about how to engage with tools like ChatGPT more critically.

“One of the things we learned hands-on toward the end of our project was to be conscious to maintain a critical distance,” she said. “And to me when we keep listening to the outputs from our prompts, it’s easy to listen to what it’s saying and you just forget that it's not a person.”

Generating creativity: AI-powered personas lead student pitches

The teams came back together the next day to continue pushing on their ideas, with the final round concluding in a pitch-style competition.

Among the ten teams emerged new personas and ideas, including how a world leader might make difficult policy decisions, how a choreographer could use AI to design a performance, and how a wartime correspondent might interpret and report on complex societal issues.

Students pitched their ideas to an esteemed panel of judges, made up by ASU faculty and staff, and an industry entrepreneur.

By the end of day two, two ideas emerged victorious. The first prize went to DanceGPT, who explored how artists and creators might use AI tools to curate an entire performance in the context of different cultures. Using Firefly, the team created a choreographer persona and an AI persona to showcase how AI could produce a technically sound dance and culturally appropriate performance.

In their final presentation, team DanceGPT noted “We discovered AI's potential in expressing dance forms, costume design, and choreographic frameworks but encountered limitations in understanding emotions.”

ASU student participant Saurabh Zinjad shares his experience in "Exploring the Depths of AI Story writing in the Prompt Engineering Hackathon".

Second place went to LLM Brew who tested the capabilities of AI to create personas for various mediums across TV and literature. Using the popular TV hit, “Friends” and focusing on the character of Gunther, the group created a cast of media experts and writers including a TV critic, TV writers, and a novelist, and a movie writer to see how each persona would bring Gunter to life in their respective mediums.

Empowering humanities students to lead in AI innovation

Oftentimes new technologies are associated with computer science or engineering students, but Maynard reminded participants that AI careers are no longer restricted to coders.

“If you think about it up until now, if you want to get a machine to do something for you, you've got to learn how to code which means you have to be in engineering or computer science,” he said.

In regards to the Hackathon, Maynard said this meant students in the humanities fields had the upper hand.

“They understand how to use language. This is transformational and it means that the students with training in how to use language, actually have the ability to develop and use systems in AI far more effectively than somebody who may have only been taught how to code.”