This week, we spoke a bit about AI and what it might look like in the coming months. We read an excerpt from “You Look Like a Thing and I Love You”, by Janelle Shane, which discussed the rudiments of machine learning, and how these algorithms could lead negative consequences. It ties back to our Weapons of Math Destruction reading a bit (I think a lot of our readings tie back to that concept). Something like ChatGPT is trained through user input; it is built upon an incredible amount of text, but ChatGPT didn’t go searching training data. That was OpenAI, who may have biases that then can show through in ChatGPT. Even last semester, using Playlab.ai with Dr. Rao’s special topics course, you could tell that each chatbot acted different simply based off of the student who developed it.

This leads to our Fortune Teller project. Another type of algorithm that could loosely be called “AI”, our Fortune Teller project included prewritten answers to questions that the user input. For my project, I went the slightly less traditional route, and made a “how far can you throw something” generator.

I wrote 14 prompts, each of which include the user’s input in the output, which makes it that much more believable that the user is speaking to an “Artificial Intelligence”. It was so much fun, and the JavaScript came relatively easily to me (after building from Dr. Whalen’s foundation.)

Leave a Reply

Your email address will not be published. Required fields are marked *

css.php