Field Test Proposal: Chatbots for #SciComm

I’m a big proponent of science communication (colloquially, #scicomm) and by extension science journalism.  I think well-written storytelling is key to getting people to understand basic science concepts that they might not otherwise have a handle on, or care about.  Books like The Immortal Life of Henrietta Lacks, The Mismeasure of Man, and Evolution for Everyone have played a large role in helping me understand the world around me.  I spent five years of higher education studying physical anthropology, and while I didn’t end up pursuing that field (very similar to skeletal biology, for those not in the know), I’ve benefited and learned a lot from science.  As a budding journalist, I want to find a way to pay that back.

The site StatNews is a daily stop for me.  They’ve been writing a lot about Zika recently – with good reason, as the disease is a national health concern – and I want to take the opportunity of my Emerging Media Platforms field test to see how a chatbot might help a site like StatNews further this kind of science reporting.  I propose a sort of “ZikaBot” – a chatbot that will answer any questions readers have about one particular disease so that they can better understand the stories they’re reading on that topic.

Consider this the basic grounding of science that readers might not have gotten in classrooms, even those who are scientifically literate, due to the recent emergence of this health concern (after they finished their studies).

Information on Zika is constantly changing as we work towards a vaccine, and an emerging media technology/platform like a chatbot is the perfect way to provide this background information to curious readers, since it can be updated on the backend whenever new information is published and will save reporters valuable time creating their stories since a basic understanding can be assumed.

This is the field test that I propose:


Chatfuel chatbot builder

Journalism hypothesis:

The existence of a chatbot that can answer basic questions about a public health concern like the Zika virus can facilitate the work of science journalists by making the public better able to understand a complicated science topic.  While reading an article on a scientific news site like StatNews, for example, they can ask their questions of the chatbot, instead of emailing the journalist, Googling and getting information from dubious (or at least not up-to-date) sources or, worse, reading the entire article and not understanding important parts of it.  The ZikaBot will provide basic clarity in questions that come up as they read the materials and will allow different people to ask different questions of the same article.

Journalism plan:

As a trial run, I will create a chatbot using Chatfuel, which will focus on different questions readers might ask about the Zika virus in particular.  The chat responses will direct them to reputable sources like the National Institute of Allergy and Infectious Disease,, and other reliable science-and-fact-based sites.  In this way, a site visitor curious about the effects of Zika on pregnancy could query the chatbot (in future, built into the article page) rather than having to leave the site to find these answers.

To test my hypothesis that this helps people understand an article better, I will circulate the chatbot along with 1 specific StatNews article, asking people to read the article and then query the chatbot with any questions they may have.  At the end of their exploration, I will ask them to rate the experience with the chatbot and how it informed their reading of the article.

If this hypothesis holds up, in future, science journalism chatbots like the ZikaBot could be created to tackle any number of topics that readers might be interested in, and used in parallel with articles on the subject.

Potential pitfalls:

The most important concern is that of human error; in building the chatbot, I might miss some key topics related to Zika.  To prevent this, I will follow key Q&As from the Center for Disease Control, World Health Organization and more to ensure that the resulting answers/topics are comprehensive.

Other concerns – responses should be in language readers can understand; after all, this is a clarification tool for complex journalism.  What if the sites linked to for explanation are just as confounding?  Rather than providing clarity to readers, we could muddle their understanding further.

Finally, there’s the concern of sending people off the site.  Much of today’s journalism is built around keeping readers on a site to collect advertising revenue from their time on page.  If we send them off the page to ask questions of a chatbot, we could decimate the revenue stream for an article.  It is my hope that in future, this type of chatbot could be embedded on the page itself.  But even then, the current model I am proposing would send readers to a second page to get their answers (since they are not handwritten by the human behind the chatbot, but rather, come from reputable sources [other sites]).

Overall, however, I think a science inquiry bot like the ZikaBot could be really helpful for laypeople interested in a scientific topic but daunted by wading into the literature or high brow reporting.  Stay tuned to this blog for updates as I build the bot, conduct my field test, and report back on whether or not my hypothesis was correct.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s