Read & Respond week 10: Chatbots

March 16, 2019

Over our next few classes, you’re going to build some bots! Chatbots – or “conversational agents” if you’re fancy – are a fairly new entry (with apologies to Eliza) into our mass communication world. Digital assistants like Siri and Alexa are chatbots: You speak to them, and they rely on rules or artificial intelligence (or both) to answer or perform a function.

Here’s a simple example: Cleverbot

A chatbot typically mimics human speech in a call-and-response way:

+ I say “Hello”

– bot responds “Hi there!”

That’s not too exciting, but bots are rarely so simple. More likely, the bot will ask what I’m looking for, shopping for, or otherwise would like to know. The Loebner Prize is a contest that seeks bots that can best approximate human interaction – as these transcripts show, they’ve still got a way to go.

Do people actually like chatbots? Yes, actually! They certainly don’t work in every situation, but the simulated human communication often goes over pretty well. The health care profession has even been exploring using bots for personal care and assisting those suffering from dementia. Language learning apps such Duolingo also have included chatbots that will converse with you in the language of your choice.

Chatbots are all around you, and they’re only going to get smarter. Although you’re most likely to see them in customer-service areas, journalists have been experimenting with them as well…

Bots are great for daily functions like answering common questions or automating data-oriented tasks or for larger projects like the one above. They’re not hard to make, either – free platforms like Dexter (which we’ll use in class) let you create simple ones on your own.

For this week’s response, reflect on what you’ve learned about chatbots and where you might apply them in journalism. I’d also like you to come up with a specific topic for a bot related to your personal or group blog. Think like you’re writing an explainer: What are some questions readers might have, and how might it respond? What kinds of questions might you NOT expect (people are weird, after all)?

Remember, your responses are due by 11:59 p.m. Monday, March 18, as a comment to this post.