SIP explores role of generative AI – 8/3/23
SIP explores role of generative AI
Students level up to find evolving industry
Draft – by Jon Cain, WPI
When WPI interactive media and game development major Kerri Thornton found out that she would be expected to consider using generative artificial intelligence (AI) during her on-campus summer internship in video game development, she was skeptical. “Honestly, my first reaction was ‘oh no! I know where this is going,’” the rising senior at WPI says, “and it’s not somewhere I want to be dipping my toe, because I know there’s a lot of controversy surrounding AI and the ethics surrounding it.”
Thornton is one of 26 college students from 16 schools spending 11 weeks at WPI as interns in the Massachusetts Digital Games Institute (MassDigi) Summer Innovation Program (SIP). MassDigi is a WPI-based center for entrepreneurship, academic cooperation, and economic development across the Massachusetts games ecosystem. SIP interns include students of computer science, game design, art, and music. Each of the four teams in the program spends the summer making a video game, from conceptualizing a prototype to presenting a final product. Along the way, they learn how to work as a team, manage project deadlines, and receive feedback and mentorship.
This year, the program’s leaders asked participants to consider using generative AI in the pre-production stage of brainstorming ideas. The technology uses deep learning to create text, programming code, images, and graphics after a person gives it prompts and parameters. Monty Sharma, Managing Director of MassDigi, asked this cohort to try out AI because companies are using it in game development. “It’s there, and the students ought to be able to use it,” Sharma says. “This is a technology that is immediately useful to a lot of people. We’re in a tech business and you need to spend your life looking forward.”
Despite her concerns about AI, Thornton used generative AI to create a graphic of a gnome as she brainstormed artwork ideas for her team’s game. She says the gnome didn’t look quite right. It was missing an arm, for one thing. Because of the limitations of generative AI, Thornton says “it seems to help with visualizing things in concept more than actual image creation.”
She thinks generative AI will become more capable and, if copyright issues are resolved, could help artists save time with repetitive tasks so they can focus on their most creative work. She hopes to try a new AI tool that starts with a pool of open-source photos and then promises to create a graphic that adjusts for changeable lighting conditions. For example, it could make a character graphic for midday sunlight in a field, and another version of the character for dusk in a forest.
Another SIP participant, James Robinson of Acadia University in Nova Scotia, used generative AI to create new code for his team’s game. He also used it to identify problems in existing code. The rising senior and computer science major says the tool saved him time by using predictive text to generate suggestions for completed code that appears repeatedly in the game. However, he says the AI makes mistakes, so it requires a knowledgeable operator. “One of the misconceptions is that AI is just doing the work of the programmer. But that’s not the case,” Robinson says. “It’s a tool programmers can utilize only if they understand the code they’re trying to make.”
As the program participants try AI, MassDigi and WPI’s IMGD program are eager to learn from the students’ experiences. Leaders with MassDigi will ask the interns for feedback on generative AI after the program ends in August. Josiah Boucher, a PhD student in IMGD who studies the ethics of generative AI, is also interviewing the SIP participants during the program about what is and is not working, and their concerns. “We haven’t fully identified from an academic research perspective the potential harms and benefits,” Boucher says of the technology. “The only thing that is certain is that it is going to change things somehow.”
Gillian Smith, director of IMGD, says faculty will use the feedback from the SIP interns as they consider how to further incorporate generative AI into their curriculum at WPI. The IMGD degree program has courses on the ethics of generative AI and its use in interactive media and games.
Smith says it’s important for WPI students to be “AI ready,” so they’re prepared to enter the workforce and to think critically about the capabilities and limits of the technology. “Being able to understand the role of generative AI requires being able to think about it from a lot of different angles,” Smith says. “What does it mean for future careers? What are the ethical and social implications of adopting these kinds of technologies? I’m excited about WPI’s IMGD program being a place where we can help people think that through.”
Thornton says this summer’s experience has made generative AI seem less scary, and she feels better prepared to interview with or work for an employer that wants someone with AI skills. She wants to work in the games industry developing character art, and she has mixed feelings about how the technology could affect the field.
“One of the great things that humans have been able to do over time is make art. I don’t think generative AI will erase people making art. I do think it’s going to change the way we make art,” Thornton says. “My main concern is that people will come to view art as less process and more product.”