BOT MACDONALD
Adjunct Professor, Designer and Advocate
A Chatbot trained on my data
This GPT has been trained on interviews, presentations, articles and classes I've taught about interaction design, games, museums, immersive experiences and futurism. As an adjunct professor I don’t get paid for time outside of class. Office hours aren’t paid and neither is grading or the the it takes to prepare classes in advance of a semester. Without Union support I can’t change any of those but I can offer 24/7 virtual office hours to my students through Bot MacDonald. Using OpenAI’s ChatGPT I build a personal GPT that is trained on all of my past syllabi, writing, project documentation, interviews and presentations about interaction and experience design. It’s my information, my background, my biases and my beliefs. I know AI makes some people uncomfortable. I’m one of those people, actually. I conduct these experiments with AI and technolgy because I believe the people who are the most critical also have the power to shape what’s possible, what ethical behavior looks like, and to craft a responsible vision for these tools. I’d rather engage then sit back and critique without having explored the possibilities.
Immediate takeaways
It’s better than I thought it would be. Is the output perfect? No. I encourage all users to validate the output. Question everything. use it as a point of engagement but remain critical and don’t cite it as research. It is, however, very good at describing the interactive design process and very much reflects my beliefs and practices as a designer. I’m comfortable with my students and clients using it and encourage follow up discussion on it’s output.
Use cases
Imagine building an institutional database of your team’s curatorial knowledge and making that available to the public, on the floor of your museum or on your website. Maybe creating a database to help preserve indigenous knowledge. Putting those side by side so visitors can get multiple perspectives on a given topic. Want to share the perspective of a famous artist or scientist?
Next Steps
Make it visual. I’ve explored GPTs and deep fakes. Now I want to create a visual, naturally conversant agent, that is trained on a specific data set so users can either type or speak directly to the avatar. I believe engaging with new technologies is a path to positive output. I don’t believe the organizations that are profiting from them have ethics and best practices in mind. So it’s up to us to figure out how to use technology responsibly.
more Chatbots
Albot Einstein
Built using OpenAI’s ChatGPT, this GPT has been trained on interviews, articles and books written by (and about) Albert Einstein. Ask it questions about physics and science. Albot has access to incredible knowledge but use best AI research practices and always verify what it shares. GPTs are tools and, like most tools, they aren’t without problems. Sometimes they don’t work as planned. In that spirit, Albot is a work in progress and I’m still adding new information and documents to feed it’s database so it can handle a greater diversity of requests with more accuracy.
Deep Frank
Deep Frank was the first GPT I built using OpenAI’s ChatGPT tools for custom GPT creation. It emulates Frank Oppenheimer's knowledge and curiosity about scientific and artistic phenomena. Be curious. Ask questions and question answers. This is an AI and, like people, it is imperfect. Be smart and always validate its responses. I’m a deep admirer of Frank Oppenheimer and truly believe he contributed to making the world a better place. I hope this humble chatbot might inspire someone to learn more about him, his work, the Exploratorium, and the incredible people who have worked there.