Presenters: Christina Fajardo, Academic Technology Specialist; Josh Weiss, Director, Digital Learning Solutions; Miroslav Suzara, PhD student, Learning Sciences & Design Technology
Recording of the session:

Presentation slides

Central Questions:

• What does the journey of building a basic AI tool look like?
• What is the difference between generative AI and generative AI building tools?
• How can novice-friendly platforms help to prototype AI tools with minimal coding involved?
• What considerations should AI builders have in mind, particularly when building for educational purposes?

Key Quotes:

This is Generative AI in an educational context. There are apps now that allow you to get an explanation as if you had an expert sitting over your shoulder, giving you a bit of tutelage. (5:25, Josh Weiss)

I’m a PhD student. A lot of the things that I have to learn and work with are lots of reading and writing…. So some of the key areas that I was trying to focus on building around [were] the process of writing, the process of reflecting, and also around reading, and I had various successes and failures along exploring these and also some really cool ideas for what I might want to do next. (15:04, Miroslav Suzara)

So I think there’s interesting things that researchers, coming from the education side, have to think really critically about before launching these [apps] out in the open. (23:52, Miroslav)

Students are using [ChatGPT and GPT-3] every day. And how can we then think about how we should design [AI tools]? What should we be building? And how can we make that as a part of the education ecosystem? (47:34, Miroslav)

Take-aways:

AI Starter Kit

• An evolving guide to help users approach their own AI building process with sets of guiding questions.
• Follows three personas throughout their building journey with example scenarios and outcomes.
• Contains a growing list of tools, including integrations with other apps and building platforms.

Identifying Pain Points and Selecting Tools 

• Avoid the “tool first” trap by identifying the problem you are trying to solve, then choose the tool(s).
• Be intentional about using generative AI to solve this problem rather than using AI because everyone else is using it.
  • Prototype it first to determine if this is something that AI should address right now.
• Keep the learner at the center of the problem you’re solving, especially in the education space

Data Security and Transparency

• Continue due diligence surrounding data collection and usage, and read the terms of service carefully.
• Understand key terms often mentioned in the fine print, such as open source (publicly accessible to use, modify, and distribute).
• Ascertain a host company’s transparency regarding their data collection when using AI building platforms.

Start Small

• Familiarize yourself with AI first before jumping into building and figure out what type of tool would work best for your context.
• Start with free (or freemium) building apps, even if they don’t meet all of your criteria, before you invest more time and money.
• Focus on satisfying one pain point and keep it simple in a one-to-one ratio before scaling up.

Tools to Help You Start Today:

No/Low Code API Connections: Zapier | n8n | AVflow
Low Code Build Platform: Buildspace
Code Required Build Platform: Codespaces

Additional Links Shared During the Workshop:

Zapier’s OpenAI Integrations
Ben Bite’s AI Newsletter
OpenAI Playground
Quickstart with OpenAI (can use NodeJS or Python repositories)