Marvista: Exploring the Design of a Human-AI Collaborative News Reading Tool

Xiang 'Anthony' Chen,Chien-Sheng Wu,Lidiya Murakhovs'ka,Philippe Laban,Tong Niu,Wenhao Liu,Caiming Xiong

DOI: https://doi.org/10.48550/arXiv.2207.08401

2023-06-24

Abstract:We explore the design of Marvista -- a human-AI collaborative tool that employs a suite of natural language processing models to provide end-to-end support for reading online news articles. Before reading an article, Marvista helps a user plan what to read by filtering text based on how much time one can spend and what questions one is interested to find out from the article. During reading, Marvista helps the user reflect on their understanding of each paragraph with AI-generated questions. After reading, Marvista generates an explainable human-AI summary that combines both AI's processing of the text, the user's reading behavior, and user-generated data in the reading process. In contrast to prior work that offered (content-independent) interaction techniques or devices for reading, Marvista takes a human-AI collaborative approach that contributes text-specific guidance (content-aware) to support the entire reading process.

Human-Computer Interaction

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the challenges that people face when reading news articles online in the digital age. With the increase in online media consumption, people may feel overwhelmed when dealing with a large amount of information. Especially in the fast and distracting way of information browsing, it is difficult to conduct in - depth reading and thinking. In addition, people with low literacy skills also have difficulties in understanding texts. Even for those who can understand texts, it may be difficult to "truly understand and appreciate the content of the texts".

To address these challenges, the paper introduces the design of Marvista - a human - machine collaborative news - reading tool that aims to provide end - to - end support through natural language processing (NLP) models. The main goal of Marvista is to help readers read online news more effectively in the following ways:

1. **Before reading**:
- **Time filter**: Allows readers to plan the content to be read according to the available time. The system will recommend the most important parts according to the time limit.
- **Question filter**: Generates a list of questions related to the article. Readers can select the questions they are interested in, and the system will highlight the text parts that can answer these questions.

2. **During reading**:
- **Focus mode**: Displays only one paragraph of text at a time and makes other paragraphs and interface elements transparent to help readers concentrate.
- **Reflection function**: Generates questions for each paragraph of text to encourage readers to reflect on their understanding of this paragraph.

3. **After reading**:
- **Human - machine collaborative summary**: Combines the article text processed by AI, the user's reading behavior and the user - generated data (such as notes and highlighted texts) to generate a personalized summary. Users can also view how each sentence is derived from a specific paragraph or user - generated data, thereby improving the interpretability of the summary.

In this way, Marvista not only provides technical support, but also promotes effective collaboration between humans and machines, helping readers navigate and understand texts more strategically, especially when facing a large amount of online information.