Until this morning, I had no idea how Twitter bots work. To be honest, they haven’t impinged too much on my consciousness, but a friend once commented to me that he uses bots to tweet stuff about media (I had no idea what he meant), and a MOOC that I’m taking uses “Teacher bots” to reply to tweets from students.
That really intrigued me, especially since the MOOC is all about the automatisation of learning (not advocating it, just asking us to think about it, and giving us lots of thoughtful readings to show us how little we really have thought about it…). The course goes into depth on, among other topics, how technology affects our definition of what it means to be human, so I’m not sure if using a teacher bot is irony, conscious tongue-in-cheek-ness, or a sinister sub-plot to see if we’re paying attention.
So, what is a Twitter bot? And what do they have to do with Teacher bots? And will Teacher bots take over education feedback?
Bots are a little computer programs that tap directly into the Twitter feeds, extracting tweets or even only parts of tweets according to criteria that you set. They can be created in pretty much any programming language (and if you want some guidance as to how, check here and here), and if you don’t have the technical experience to do it, a developer could whip one up for you in a few minutes.
According to Twitter, almost 10% of its active accounts are automated bots, which means, doing the math, that there are probably over 20 million automated Twitter tweeters out there. What on earth are they tweeting about?
About half of the bots are spam, trying to sell followers, trying to manipulate the trending topics, or trying to sell you body parts enhancers. Twitter has a strict anti-spam policy, and spam accounts are eliminated once caught. Catching them is not that easy, unfortunately.
But the majority of the bots are simple re-tweeters. I noticed the other day, for example, that anything that I tweet with the hashtag #startup gets retweeted by another account. It doesn’t do that because thinks my tweets are particularly insightful (although they are, of course, #humblebrag). It does that because it’s programmed to. And those kind of bots can be very useful.
You can program a bot to retweet any tweet with certain words in it, hashtags or not. Companies are increasingly using this tactic to find out what people are saying about them. You’ve probably noticed it when you tweet something nice about a particular brand and it’s suddenly retweeted by the brand’s account. Or that if you complain about a service or product, someone from the Customer Experience department gets in touch with you right away. Media outlets are increasingly using this technique to find sources and information. Their bots retweet messages that include certain keywords, such as “Ferguson” or “immigration”, which feeds them both material and potential collaborators.
As with the media outlets, bots can also help students and teachers to find information and sources, for use in class debate or in essay preparation. They provide up-to-date, real time information which can be incorporated into presentations and assignments, giving a different and more emotional perspective than text books or even prepared videos.
Programming bots to automatically retweet is pretty simple. However, we are beginning to see a rise in the number of bots that attempt “dialogue”, using a mixture of algorithms and basic artificial intelligence (AI). And here’s where its potential impact on education gets interesting…
For now, the AI bots are still clunky. Tweets like “That’s a very good comment, Hank” and “Why are you saying this now, and how can I help?” aren’t really going to fool anyone for long. But as we get more experience with this, and as we learn more about AI programming, we will soon be receiving messages that seem almost… human.
So, is that a real-live human tweeting a reaction, or a bot? Is that your teacher responding to your query, or a computer program? Does it matter?
Yes, it does, very much. It’s a question of trust.
I love the efficiency of Twitter communication, and I strongly believe that it should be used more and more in education. And I believe that anything that can be automated should be automated, to free up time to do more meaningful things. I also understand that the role of the teacher in education is being re-examined, with the advent of MOOCs and flipped classrooms and excellent teaching apps for the tablets that more and more students are using.
But no one can deny the value of the influence of a good mentor. Since the beginning of documented history, students have had figures to look up to and to learn from. Since the beginning of documented history, certain eloquent and intelligent individuals have taught, inspired and motivated the younger generation. That has not changed, even though the methods of delivering a large chunk of the information that we believe the students should learn, has. Ask any young person who their favourite teacher has been so far, and watch their faces light up as they remember how doors were opened, lights were turned on, and a whole lot of other appropriate metaphors. Their faces light up because the sensation of understanding and of seeing a future roll out before you is indescribable, and one of the best gifts we can give the next generation.
Can an automated response do that? The question is actually a very big one, and brings up a ton of artificial intelligence issues. Can we love a robot? Yes, it seems that we can, if they have enough human characteristics (tell me you didn’t cry in Wall-E, or I Robot). But loving something doesn’t mean that we are going to trust it enough to let it inspire us.
Can we count on a robot to care about us? Here’s where things get sticky. Any good teacher has no problem showing that they care about their students, as people. They feel proud when the children exceed their own expectations, they feel happy when the class is excited, they understand if someone is having difficulties. Good teaching is not about imparting knowledge or grading papers, a robot can do that. It’s about feelings. If we feel that someone cares about us, we are more “invested”, we try harder, we want the teacher to feel that his or her caring is not misplaced. It’s natural, and not just in children. (Of course there are exceptions, and they have many posts dedicated to them in the psychology blogs.)
As the saying goes, “You can fool all of the people some of the time, and some of the people all of the time, but you can’t fool all of the people all of the time.” A good AI bot can probably fake it to some extent, but nowhere near enough (yet) to convince us that it’s human. And as for connecting, inspiring, and showing that they care… Can we trust them to do that? It’s also often hard to trust humans, but with humans we have the emotional connection of forgiveness, understanding and affection. The human presence allows for a wide range of non-verbal communication as well, all of which generates the trust needed to really inspire another human.
Is the human presence necessary for learning? No, not for all learning. I can learn from reading books and watching documentaries, and while they were admittedly created by humans, there isn’t a direct human presence. And I can learn from Google search results, I can learn from a walk in the woods in which I observe moss patterns.
Is the human presence necessary for teaching? Yes. While videos, books, online courses are very informative and even instructive, they are limited to “showing”, which is not the same as teaching. Valuable, but not the same. For a good teacher, you need the human touch.
The MOOC that I mentioned at the beginning of this post managed a powerful mix of the human and the automated. The classes were via videos and readings, and there was an active discussion board, as with most other MOOCs. What set this class apart from others is that they also organized Hangouts, live video chats in which students could participate. Participation is limited to 10, but they had more than one session, and the chance to “meet” both the professors and other students made it feel so much more real, important, even meaningful. It’s not the same as a physical meeting, but it’s close enough. It feels personal.
For the course, as an experiment, the professors created a Teacher Bot, which responded to students’ tweets with responses according to recognizable words in the tweets. It has its usefulness, no doubt. But its answers were not human, and so did not carry any “weight”. There was a lack of trust, and engagement. Tweets are useful for receiving information. But that can also be taken care of on a static web page. Twitter is even more useful for connecting, however briefly, with someone else, to exchange thoughts, jokes, ideas and links. While there are many hilarious bot accounts (I’ll tell you about some of them in another post), and I confess that I subscribe to some of them, Twitter’s value is in the human connections. Bots can be worthwhile, and even entertaining. But emotions and inspiration need a human source.
The bots put up a good effort, though. I ended up feeling a touch of affection for our Teacher Bot. His tweets were getting stranger and stranger. He seemed to give up trying to appear human. Although, the last tweet I got from him made me wonder, and almost brought tears to my eyes:
— x —
For more information on Twitter, check out my Flipboard Twitter magazine: