This in-presentation activity will help you explore some of the trials and triumphs of AI-incorporation into business!

1. Get into groups of 2 - 4 for this activity.

2. Elect ONE group member to be the recorder, who will need to create several slides documenting the group's work.

3. Skim each of the activity Options described below and pick ONE that your group will discuss and record!

A few preliminaries before we begin:

1. Grab a copy of the Google Slide templates here:

   Option 1 Presentation Template (Google Slides)

   
   

2. Click the Use Template button at the top of the slides to begin your own copy.

3. Follow the next steps to create several (simple) slides to document your findings related to the discussion's material!

There will be no obligation to present your slides, but I'll ask for volunteers to share in the final minutes!

Note also: although the following descriptions mention 1-slide each, you're free to add others if you see fit.

Feel free to skip any slide if you're more interested in what follows!

Data Dilemma

Get practice finding datasets that might be amenable to machine learning for some business task OR propose how a generative model like ChatGPT or Midjourney could be used to fabricate that data!

1. Go to a data repository from amongst those listed on the Wiki page, or just jump straight to either Kaggle or GitHub's Awesome Public Datasets:

   Wiki - ML Datasets

   
   

   Kaggle Datasets

   
   

   GitHub Awesome Datasets

   
   

2. Find a dataset that you find interesting and could envision becoming useful to an app or business through the tools of ML.

3. Take a screenshot of the dataset or its description and add that to Slide 1.

4. Lastly, on Slide 1, briefly describe your business idea surrounding the dataset in terms of inputs and outputs of a model fit to it.

Model Search

Get practice finding pre-trained models that might be used in some business task OR propose an interesting use-case for a generative model like ChatGPT or Midjourney!

1. Go to HuggingFace's repository of public models and explore a category of interest, e.g., image-to-text or text classification:

   HuggingFace

   
   

   Hint: click the Inference Available button at the top to see examples that you can test right on the site!

2. Find a model that you find interesting and could envision becoming useful to an app or business.

3. Take a screenshot of the model or its description and add that to Slide 2.

4. Lastly, on Slide 2, briefly describe your business idea surrounding the model. This may or may not be consistent with your dataset found on Slide 1.

Not sure what you're looking for?

Here's an example from the slides:

• Check out a simple image classifier that takes in an image of a person and tries to guess their age:

  Age Classifier

  
  

• In the right panel, find the "Inference Provider" to play around with the model to see what it can do!

• As for an accompanying business idea, consider an app that uses GenAI to provide text chat from some character shown in an image with language consistent with their age!

AI in Apps

Now that you've seen some data and models, time to be creative!

This one's wide open!

On your final Slide 3, give a lightning pitch for an AI-enabled app or business that incorporates some of the data or model ideas from the lecture or your previous slides

A few preliminaries before we begin:

1. Grab a copy of the Google Slide templates here:

   Option 2 Presentation Template (Google Slides)

   
   

2. Click the Use Template button at the top of the slides to begin your own copy.

3. Follow the next steps to create several (simple) slides to document your findings related to the discussion's material!

There will be no obligation to present your slides, but I'll ask for volunteers to share in the final minutes!

Note also: although the following descriptions mention 1-slide each, you're free to add others if you see fit.

Feel free to skip any slide if you're more interested in what follows!

AI Gone Awry

Find an example of a company or product that either failed or risk failing because of their misuse-of or inability to obtain a ROI through GenAI.

1. Search Google, ask ChatGPT, or even highlight a known story of a failed venture -- as long as it involves GenAI, it's perfect for this activity.

   Try to find a case-study in a domain of your team's interest, e.g., if you're interested in EduTech, find a story of a business that used GenAI but failed in this sector.

2. Once you have found such a case-study, summarize their intended product and what happened to them on your first slide.

Diagnosis and Prognosis

Time to dive into what went wrong with the business you found and make recommendations to avoid similar pitfalls!

1. Analyze what went wrong with the company's application of GenAI. You may want to focus on some of the GenAI Do's and Don'ts discussed earlier in the presentation.

2. On your second slide, include both (1) a diagnosis of *why* GenAI was not the right tool for the job / failed to obtain a ROI and (2) propose an alternative or suggest another way that the company might have tried to approach its business model.

Adjacent Possible

Now's your chance for some creativity! Propose your own GenAI-assisted startup that is *similar* but *better* than / *fixes* the issues of your case-study business.

1. Lots of room for creativity here, nothing needs to be fully-thought-out for it to generate discussion!

2. Include your brainstorming / elevator-pitch on your 3rd slide.

This one's easy -- when you're done with your slides, make sure to share them and then place a link in the following document:

BEST Slide Links