Chatbots are flavour of the month for tech startups and giants alike. They're a natural extension of the search engine - all the world's knowledge delivered conversationally on messaging platforms that already consume much of our connected lives. What seems like a discussion with a person is actually a conversation with a computer that responds automatically to human input. Facebook is investing heavily in chatbots. So is Google.
Chatbots will have already arrived in the lives of students (they've been appearing in Facebook Messenger and WhatsApp for some time) but what do teachers need to know? Should you be thrilled or fearful about the prospect of bots in the classroom?
Let's establish something first of all: bots aren't that clever yet. Here's part of a conversation I had with Cleverbot:
Me: What is your name?
Cleverbot: Carlos. What's your name?
Me: Will.
Cleverbot: Will is a Japanese name?
Me: No, it's short for William. William is a Germanic name.
Cleverbot: What is a nickname for Alexandra?
Me: Alex.
Cleverbot: No it isn't. Your name is Alex.
Me: No, my name is Will.
Cleverbot: Cool, that is my name too.
Me: No, your name is Carlos.
Cleverbot: Your name is Reginald Loudermilk.
Me: No it isn't!
If you teach pre-schoolers, this kind of conversation may have a familiar ring to it. But it certainly won't make you fear for your job. How it works, however, does present real opportunities for teachers and students.
Broadly speaking there are two kinds of bot - those that follow 'rules' and those that 'learn'. To see a rule-based bot in action take a look at the one I set up for the purposes of this post. Ask the bot what 2 + 2 is and it will give you the correct answer. But ask it what 2 + 3 is and it doesn't know. That's because it was programmed to respond to the written phrase '2 + 2' with the response '4'.
I could have programmed it to respond to a wider range or questions (actually, you can ask it a few other things - give it a try) but the point is that somebody has to write these rules - the chat bot has no inherent knowledge. That's why many bots prompt you to use particular phrases. It's the equivalent of being asked to press 1 to hear your balance or 2 to make a transfer when you're on the phone to your bank. And who doesn't enjoy that?
A more elaborate rule-based bot might feed off a much larger data set. For example, if you connected a bot up to Wikipedia you could program it to respond to commonly formulated questions like 'when was...' or 'who is...' with appropriate information. Or imagine it connected to Wolfram Alpha, a service that can already offer meaningful answers, next steps, and supplementary information when asked questions like 'what is the Battle of Hastings' or 'what is the square root of 67567899986677654444332'. If you have an Apple device you might have noticed that Siri relies on Wolfram Alpha for some of it's answers - this stuff is happening right now.
Cleverbot, the not-so-clever bot I was chatting with earlier, is a bot that learns. Although it hasn't quite mastered the art of conversation yet it's becoming smarter every day. Every chat it has helps it to adapt to the next. It still has a lot of learning to do but once it understands how it is meant to respond and has enough data available it will become a powerful way to obtain information using natural language.
Educational applications for such technologies are obvious. Students could quickly obtain a mathematical formula or the date of a battle without the need of a teacher to remind them. Analysis of the kinds of questions students ask could guide teaching style or highlight knowledge gaps.
Bots may eventually be able to teach a concept, stopping to test knowledge throughout the chat and offering supplementary information where required. Critically, the platform for delivery will be familiar, friendly and human enough to be engaging.
A concern for many educators will be the quality of responses. Before you hand over the teaching to machines you need to trust that the underlying data is reliable. Many bots will rely on open source data sets like Wikipedia and some educators will be rightly suspicious about this. Close collaboration between teachers and developers will be critical if bots are going to win the trust they need to play a useful educational role.
And what of the human element? Can a bot inspire? Will a bot know when a student has had an argument with her best friend? Artificial Intelligence is coming to the classroom, but Artificial Emotional Intelligence? That's a long way off. A great teacher can't be replaced by technology but they can be assisted. That's where the opportunity lies: in giving teachers and parents the best technologies to support learning, without losing the essential human touch.