Exploring the Potential Use of Generative AI Technology from Video Games to Academia
By Ye Hwa Kang, News & Campus Life Staff Writer
On September 25, 2023, OpenAI announced that ChatGPT will be able to “see, hear, and speak.” OpenAI went on to reveal that ChatGPT Plus users will be granted access to the new features on iOS and Android. The voices and fluency of conversation showcased on this organization’s website were so human-like that one would not have believed that it was all AI-generated.
Generative AI’s potential to speak, hear, and see is strong, especially when it comes to video games and virtual reality. NPCs (non-playable characters) will be free from their scripted lines and move on to having vocal conversations with players. In fact, several games have already been released that utilize this exact feature—including Inworld Origins.
In Inworld Origins, the player acts as a detective to solve a mystery case by conversing with the NPCs. Speaking from my own experience, I was blown away by the game’s integration of conversational skills that allowed players to converse freely, in addition to the scripted features. For example, if I bring up something irrelevant to the story’s progress, the characters would acknowledge my comment but state that it was unimportant to the storyline or conversation. It is only a matter of time before more video and virtual reality games incorporate AI NPCs, as they have tremendous potential for entertainment.
AI NPCs can also be implemented in existing games through mods (short for ‘modification’—a program that allows players to modify existing games to their taste). One example would be the implementation of Inworld AI mod to the Elder Scrolls V: Skyrim, a game developed by Bethesda Games Studios. Skyrim typically just provides scripted NPC dialogues, but one user gave voices to the NPCs and allowed players to have a conversation with them through the use of Inworld AI mod. The mod retained the personalities of the NPCs, which hints to us on how video games might look in the future with generative AI.
Although generative AI has been relatively well received in the field of entertainment, it has caused trouble in academia. More specifically, there have been cases of students using generative AI to do their assignments for them, which is an act of plagiarism and a violation of academic integrity. A UNESCO global survey suggests that the majority of schools around the globe have yet to create school policies regarding generative AI use by students. That being said, U of T is one of the few schools that have such institutional policies in place. In April 2023, the university provided sample statements regarding the use of generative AI technologies such as ChatGPT to include in their course syllabi. A U of T website that teaches students about Academic Integrity states that if students fail to comply with such regulations, they will have violated the University’s Code of Behaviour on Academic Matters.
Although generative AI technology is usually seen as an obstruction to academia, could there be a way for U of T to use this technology for the betterment of faculty and students? Perhaps generative AI can be implemented in U of T library chats after hours to provide additional support. This would be an effective use of this technology given that the latest time the library chat closes is at 10 p.m., and many students work on assignments late at night and need assistance after library operating hours. In cases like this, it would be helpful for students in need of assistance to be directed to the U of T library chatbot that would give further information and instruction. In a general sense, it would be more convenient to have a U of T chatbot that answers students’ administrative questions instead of students having to move from one website to another to search for the information.
C, a first-year U of T student who wished to stay anonymous, expressed her thoughts on the use of generative AI in school websites. “I haven’t used too much of ChatGPT. I think it’s definitely a good resource. I know that some of my friends have used it to like, summarize articles and stuff for them, and I think some of my disabled friends have used it to make school a more accessible process, especially when they couldn’t find accessibility within their own schooling. For me, I think it’s just kind of like a fun tool to play with. I don’t really think it has a place for the most part in academia, for like actual assignments.”
When asked to consider the potential use of generative AI technology as chatbots on U of T websites, C mentioned that “the admissions process having that integrated into U of T would be really helpful . . . I mean, [applicants] can reach out by email, but if you’re getting hundreds of applicants, and you are a singular person, it’s kind of hard to respond to all of those emails, and U of T has so many different layers of websites.”
C also stated it would be easier “if there was a system where you could be like ‘do you have this book in stock’ and just have something easily answer . . . it could be a good way for students to find information quickly.”
Indeed, U of T has numerous websites that cover different information for each of the three campuses. Given the vast amount of information, it would be more efficient and intuitive for students to search through a chatbot that uses generative AI instead of emailing the department to ask questions about information that is not listed on the website.
All of this being said, C seemed pessimistic about the correctness of generative AI technology. She states, “I do worry though, about the idea that it would replace staff because I definitely don’t think that’s possible. I think that where AI is right now, it can’t replace human beings, and I definitely think there’s issues with AI and even ChatGPT; like sometimes it will just get into loops where it will respond the same thing over and over again . . . I think it’s definitely important to keep the human element of it.”
As C points out, it would be crucial for U of T to ensure that accurate information is provided to students on school websites if implementing generative AI technology. While generative AI is only beginning to gain momentum in the technology industry, it cannot be fully trusted yet to be implemented on U of T websites if there is a high chance it could deliver misinformation.
Nevertheless, generative AI is here to stay. In the coming decades, U of T websites may have this feature in the future, minus the misinformation.