david warren – Extension Foundation

Read the full report here.

Summary

U.S. Cooperative Extension provides useful, relevant, local, research-based, information to citizens throughout the country. Modern affordances like web searches, email, and instant messaging have built the expectation that information and services are available in real time and via the Internet–today’s primary source of information.

The local, nuanced, time-sensitive information that Extension and the Land Grant Universities provide exists, but can be difficult to find. Even more, the best resources are currently distributed across many data sources: e.g. Land Grant University information pages, Cooperative Extension websites and databases, Regional Integrated Pest Management Centers, Regional Rural Development Centers, and even eXtension’s massive “Ask an Expert” database of questions and answers.

This investigation looks at one solution to the problem of information fragmentation and “findability” of research-based answers with local relevance: the creation of a distributed Ask Extension data registry and search interfaces. This solution would pull together data sources from throughout the Extension ecosystem to answer questions directly and accurately and connect local citizens with their local Extension professionals, products, and services.

Advances in technology such as machine learning, Artificial Intelligence (AI), and related data tools present the possibility of capitalizing upon this opportunity, by offering smart and localized access to the addressable research-based information. In addition, they offer long-term scalability that will answer common questions immediately, providing 24/7/365 availability of Extension resources.

Read the full report here.

By David Warren
eXtension Artificial Intelligence Fellow

Alexa, Siri, Google Search and other consumer-facing machine learning services have revolutionized the way that many people find answers to questions that occur in everyday life. With this thought in mind, eXtension is investigating how AI and machine learning tools could be applied to both improve access to information and increase the local impact of Extension Services.

Our investigation has led to a scoping effort and preparation of a grant application that envisions the creation of a machine learning service. This service would accept questions from clients of partner Extension Services via text input, and deliver answers and resource links.

The type of service we are envisioning is a “chatbot”, which will utilize machine learning algorithms to find the best answers to questions. The chatbot will built on top of eXtension’s large database of thousands of questions and answers gleaned from years of the Ask an Expert service. Additional content hosted by partner Extension Services such as fact sheets, reports, links to workshops, and connections to local offices and resources could also form part of the answers that would be supplied to questioners.

The chatbot would be accessible via API’s by the various Extension Services for use with websites, text response from phones, and potentially other uses. We hope to superpower local extension websites with this service, with eXtension remaining invisible behind the scenes. The goal is to provide local information and local connections, capitalizing on national data for machine learning to provide high-quality chats/answers.

This project in no way conceived as a replacement for the vitally important people who work in Extension Services. Instead, it is intended to leverage the expertise of Extension personnel in a new way, in order to reach people who likely would not have connected to Extension. We will work hard to give it the Extension feel, with right touch and local feeling.

The initial meeting of an ad hoc advisory group was held on April 10, 2019. Attendees from Land Grant Universities included: Becky Griffin of the University of Georgia; Rose Hayden-Smith of the University of California; Steve Judd of the University of New Hampshire; Robin Baumgartner of the University of Idaho; Eli Sagor of the University of Minnesota; and David Warren of Oklahoma State University.

A second meeting of the ad hoc advisory group is planned for April 29, 2019. The grant application for this project will be submitted at the beginning of June, and if funded, the project should begin late in the summer. Look for further updates here at eXtension.org.