I follow a bunch of different people on Twitter. Some are real people I’ve met at conferences who say and share interesting things. Besides these people, I also follow topical Twitter users who share content on specific areas of interest. For instance, I started following a Twitter user that shared stuff on Apple Watches. After buying an Apple Watch last fall, I wanted to learn more about the device and thought that Twitter would be a great way to get started.
After following this Apple Watch user for a few months, I started noticing some odd behavior. First, the user shares a lot. It tweets something new almost every hour, which is a little excessive. Also, it sometimes shares the same posts over and over. If I scroll down through my Twitter feed, I’ll see that this user will share the same post several times over the course of a day. While I don’t know for sure, I’m starting to suspect that the Twitter user isn’t a human at all. After examining the user’s tweets and actions, I’m starting to think that I may be following the machinist musings of a bot.
Bots are scripts that run automated processes over the Internet. Bots can be simple or really complex. For instance, the Apple Watch Twitter bot may just be programmed to retweet anything with an Apple Watch hashtag. That would explain some of its excessive and repetitive behavior. More complicated bots could be programmed to write news articles for web publications. Sound farfetched? Check out this article from Wired magazine. Bots are already generating the articles on some major news services. Who knows? Maybe you’ve read an article or two that were written by a computer application and you didn’t even know it.
This blog is dedicated to teaching, learning and technology. So, what are the instructional applications for bots? Before anyone worries about instructors being replaced by Matrix-like computer agents who lecture and assess students, I don’t see bots taking over the instructional landscape. Like any instructional technology, bots will play a role in supporting the work that instructors do but not replace it. In a book I reviewed a few months ago, Zhao refers to the process as “dancing with robots.” Instructors will still take the lead in creatively designing lessons and assessments. Bots will complement our pedagogical work. Maybe bots will help us assess the grammar in papers. Or maybe bots will allow us to email struggling students or offer assistance to students who have missed class for several days.
But this isn’t some distant vision. This is already happening to a degree. Take this post from W. Ian O’Bryne. He used a bot in a MOOC he regularly helps to facilitate. Because the MOOC is large and participants share a lot of their work through Twitter, some students are overwhelmed by the shear volume shared by their MOOC-mates. Some students also feel isolated if nobody in the MOOC shares, likes or responds to their work. To address this, O’Bryne developed a Twitter bot that provided feedback to people when they shared MOOC-based content. He’s still tweaking the bot’s performance to add some randomness and more authenticity but he writes that he can visualize a future where bots become “concierges to learning experiences.”
The development trajectory of bots will not be without some bumps, however. When we reduce complex processes to algorithms, crazy things can happen, especially when the programs are designed to adapt and evolve in the wild. Check out this story of an algorithm that mistakenly tagged photos uploaded through Google. Or this story of the Microsoft-developed chatbot that became racist after interacting with other Twitter users. Ultimately, these issues will be worked out and better bots will be developed. Until then, however, I’m happy to say that this blog post was entirely written by a genuine human being.
Or was it…