A team of researchers from the College of Arts & Letters has released a new web application aimed at keeping online conversations on track. 

The Faciloscope app, developed by faculty and students in MSU’s Writing, Information and Digital Experience (WIDE) research center, provides rapid, real-time analysis of how online conversations are developing. This can be especially helpful to moderators, or conversation facilitators, giving them the tools they need to keep conversations moving along in a productive way or the insight to know when a conversation is not worth continuing.

All too often, online conversations, like those that take place within informal learning and discussion environments, such as Internet forums, a comments section, or social media, stray off topic and devolve into unproductive exchanges that may even become abusive, inciting personal attacks and name calling.

We are not looking to replace the nuance of human conversation with automated robot responses, and we are not trying to replace anyone’s judgment.

BILL HART-DAVIDSON

That’s where Faciloscope, a new kind of comment analysis technology, can help and even may be useful in detecting trolling behavior.

How It Works

To use the app, one must simply copy a conversation and paste it, in plain text, into the text box of the app and click the submit button.

Faciloscope then reads the text and looks for “moves” that people make that effect the overall dynamic of a conversation – moves that can keep it going or shut it down.

The conversation is broken down, using a machine-like algorithm trained to recognize the three basic functional moves participants make that move conversations along:

• Staging – an attempt to establish the ground rules of a conversation or to help others understand the social circumstances that give rise to the exchange.
• Inviting – a direct request to participate in the conversation in some way.
• Evoking – an attempt to connect ideas, often by referring to something said previously. These moves can highlight agreement or disagreement. They also can be positive or negative in terms of tone and sentiment.

“The Faciloscope takes the text of the conversation and produces a visualization of what facilitation moves are happening in that conversation,” said Bill Hart-Davidson, WIDE Senior Researcher and Associate Dean for Graduate Education for the College of Arts & Letters, who is a member of the Faciloscope research team. “But what’s important to understand is that it doesn’t evaluate any of the substance of the discussion. It’s really only looking for the structure of the conversation that tends to keep it moving, that tends to invite other people in, and that tends to make connections between what people are saying.

“In a way, the Faciloscope is paying attention to those little cues that we sometimes take for granted as humans but that are really important in forming the fabric of a conversation, or a deliberative exchange.”

Multiple Uses

Originally developed as a tool for science museums in partnership with the Science Museum of Minnesota and the Museum of Life and Science in North Carolina, Faciloscope is part of a larger WIDE research center project, funded by the Institute of Museum and Library Services, that focuses on enhancing techniques for informal learning in science.

However, as word of the app has spread, it has become clear that the tool can serve a wider purpose.

One organization that has shown early interest in the app is SpeakUp NC, a free online tool that sends subscribers the most talked about news stories by their interest areas, allowing them to quickly and easily post comments to articles that matter most to them or to their organization. The North Carolina-based organization is exploring how Faciloscope can foster better discussions and public interaction on sites operated by newspapers, civic organizations, and the like.

We hope to demonstrate that we can do smart things with machine learning and robots of various kinds that help humans achieve human goals.

BILL HART-DAVIDSON

“What SpeakUp NC and our partner collaborators have in common is they all try to use the Internet and online forums to raise the level of public engagement and have meaningful conversations without devolving into ugly shouting matches or abusive language,” Hart-Davidson said.

Research Team

Besides Hart-Davidson, the other members of the Faciloscope research team include:

• Ryan Omizo, who led the app development as a post-doc researcher in WIDE and a fixed-term faculty member in the Department of Writing, Rhetoric, and American Cultures (WRAC). Omizo is now an Assistant Professor in the Department of Writing and Rhetoric at the University of Rhode Island.
• Minh-Tam Nguyen, WRAC Ph.D. student and WIDE Research Assistant
• Ian Clark, Undergraduate Research Assistant, who has since graduated with a major in Experience Architecture from the College of Arts & Letters. He now works as a User Experience Analyst at Auto-Owners Insurance.

Members of the team say they are not trying to replace any human effort with this app, they are just trying to augment human effort.

“We are not looking to replace the nuance of human conversation with automated robot responses, and we are not trying to replace anyone’s judgment,” Hart-Davidson said. “We hope to demonstrate that we can do smart things with machine learning and robots of various kinds that help humans achieve human goals.”