Do you feel like the more you learn about Artificial Intelligence (AI) maybe the less you feel you know? We think the best way into it is to experience it yourself, hence our previous set of activities looking at things that look real but are not.
Let’s go deeper. This week I will show you how to create an AI driven bot that can complete sentences based upon an algorithm built from your own tweets. I just made one, and I can have it complete sentences that come from my own tweets, but I never tweeted.
Here, let my bot do the talking.
When completed, you can use your bot to complete any sentence.
It takes a few steps, but they are clearly explained in the Hugging Tweets demo. You identify the twitter account to use (try your own), then it runs a bunch of code that flies by. It will download enough of your tweets to build your custom AI bot.
As a trial you can launch it with a phrase to complete. But once it is done, it creates a link that can be used (shared too) for running any other sample completion. You can try mine but this is more interesting if you create your own bot that generates tweets base don your own history.
Many will be nonsense, but others do feel real. Or do they?
So what might happen if all #netnarr students create one of these, and shares the links? This could be used to generate a narrative conversation. Like if you started with one person’s Hugging Tweet engine, with “The Future University looks like”
We could compare responses, but what if we made en entire narrative thread? What kind of fiction can you make here?
More AI to Explore
I found a collection of other types of AI you can try directly that generate text, images, even sound. The one I played with was Even Stranger which lets you upload a photo, it tries to identify what it is, and it generates a poster in the style of the Stranger Things show. It did pretty well to know that my photo of Felix is dog.
But it took another photo of an old car, actually an El Camino, that the AI identified as a “tire”
AI is.. stranger than– what? What can you learn by experimenting with AI devices? What does it mean? And what about the impact of feeding these algorithms, what questions do you have?