Generative AI in action
At a recent MIT symposium, participants got a sneak peek into 12 cutting-edge generative AI projects.
By Sara Feijo and Katherine Ouellette
Imagine a world where students have access to an application that helps them improve their public speaking skills. A world where anyone can create a mobile app that harnesses the power of generative AI — even if they don’t have programming or coding experience. That world already exists.
More than 250 people got a behind-the-scenes look at some of MIT’s cutting-edge artificial intelligence projects during the “Generative AI + Education’’ symposium. Hosted by MIT Open Learning, the symposium featured 12 interactive demonstration stations, as well as engaging discussions with education leaders who explored the future of generative AI in and beyond the classroom.
Valfee: Helping students with public speaking
Surrounded by a crowd of people, Nwanacho Nwana SB ’20 demonstrated his mobile application, Valfee, which uses AI-assisted observational learning to help students reduce their public speaking anxiety. Valfee asks students to select a role model whom they’d like to emulate as a way to practice their public speaking skills. Students then choose prompts and answer questions about those topics, and the app compares their speeches against the verbal patterns of their role models.
“The app gives feedback on how to speak slower or faster, how to reduce filler words, or how to use fewer pauses or more pauses,” Nwana explained enthusiastically. “We’re excited to hopefully change students’ lives.”
Nwana received a $27,260 education research grant this past summer from the MIT Integrated Learning Initiative to assess Valfee’s effectiveness at reducing public speaking anxiety. He is conducting that research with Lori Breslow, senior lecturer in managerial communication at MIT Sloan.
MIT Personal Robots group: Helping students learn and flourish
At another station, Research Assistant Isabella Pu eagerly spoke about the creative side of generative AI for K-12 students. Along with Safinah Ali and other members of the MIT Personal Robots group, Pu is working on middle school curricula that introduce students to AI, tools that generate creative and emotional images, and the hidden biases within these tools.
“The next generation of students is growing up among these technologies that we couldn’t even dream of not too long ago,” Pu said. “So, it’s really important to teach them about AI and how AI can help support their creative process.” Pu envisioned educators preparing students to solve the world’s problems by leveraging these tools in bold innovative ways.
Members of the Personal Robots group are also working on a storybook co-creation tool that allows students to use generative AI to write and illustrate their stories. In a story-writing activity, AI suggestions are used to inspire students to stretch their imagination and take their stories in new directions. “These curricula and tools are designed with educators in mind, come with educator guides, and are used by teachers in the classroom,” Pu said.
Another demonstration by Brayden Zhang, a research assistant at the Personal Robots group, highlighted a project that is leveraging the planning ability of language models to expand social affordances that promote personal growth. For example, the group is investigating how AI can be used to help at-risk kindergarten-age children learn language and literacy skills.
MIT App Inventor: Harnessing the power of generative AI
David Kim, a software developer for MIT App Inventor, challenged participants to think about what sorts of mobile apps students can create with these powerful generative AI tools. “Imagine a kid creating a mobile application where they list all the ingredients they have in the refrigerator and then ask ChatGPT, ‘What should I cook for dinner?’ And ChatGPT comes back suggesting a dish,” said Kim. “Imagine a kid creating a mobile application where they can generate new art based on a description of a dream they had last night.”
With MIT App Inventor, anyone can create a mobile application that harnesses the power of generative AI. Kim demonstrated how users can use App Inventor’s drag-and-drop, blocks-based programming to create applications.
Aptly: Turning ideas into working apps
At a nearby table, developer Evan Patton showcased how Aptly builds on App Inventor’s 15-year experience in democratizing mobile application creation for people without a background in coding or technology. The Aptly tool transforms spoken descriptions into working applications. For example, a user verbally instructs a computer to create an app with a text box, a list of six languages, and a button that translates text into the selected language. The result is a working app that does exactly that.
“I may program it in Italian just to show that English might not necessarily be the programming language of the future, but your own native language might be the programming language of the future,” Patton said.
Other demonstration stations featured:
- An application that gives users a more robust search experience on MIT OpenCourseWare. With natural language processing working in tandem with basic text-based searches, the app can better refine which content results are yielded for the user. Part of MIT Open Learning, OpenCourseWare offers free, online educational resources from more than 2,500 MIT courses.
- A chatbot that is an expert teacher assistant for STEM courses. Chattutor, which stemmed from the MIT Quantum Photonics and AI group, accurately answers students’ questions and can be incorporated into courses.
- A project that is investigating the use of AI-generated characters for supporting personalized learning, and its impact on learning motivation.
- Research from the MIT Fluid Interfaces research group that is looking into how generative AI technology can help people develop critical thinking skills in a world where devices put information at people’s fingertips.
- A personalized educational robot that is currently being developed by the MIT Clinical Machine Learning group. The robot will be able to help students and educators more efficiently prepare class materials.
- Research from the Fluid Interfaces group that is looking at how cutting-edge technologies, such as generative AI, can be used to tailor learning experiences in and outside of the traditional classroom.
- Open-AI-mer, a chatbot that allows users to talk with a virtual version of theoretical physicist J. Robert Oppenheimer. Open-AI-mer, which talks like Oppenheimer and knows about his papers and experiences, is part of a larger project that aims to showcase the personalities and stories of scientists who have been historically excluded from science because of the history of the field.
The “Generative AI + Education’’ symposium was part of MIT Generative AI Week. This three-day event series explored the latest cutting-edge research, the implications and possibilities of generative AI, and the opportunities and challenges posed by this technology in education, health, climate science, and management.
Visit the MIT News website for the full coverage of MIT Generative AI Week, including an overview of the entire week, a recap of the “Generative AI: Shaping the Future” symposium and a presentation on generative AI-aided art.
