CounterQuill: Investigating the Potential of Human-AI Collaboration in Online Counterspeech Writing
Project Overview
The document explores the innovative application of generative AI in education through the platform CounterQuill, which helps users develop empathetic counterspeech against online hate speech. This AI-driven tool utilizes a human-centered approach, guiding users through a structured process that includes learning about hate speech, brainstorming effective counterspeech strategies, and collaborating with AI to co-write responses. A user study demonstrated that CounterQuill significantly enhances users' confidence and comprehension regarding hate speech, while also promoting a sense of authorship and ownership in their writing. Overall, the findings suggest that generative AI can play a crucial role in education by empowering individuals to engage thoughtfully in online discourse and equipping them with the skills to navigate and counter hate speech effectively.
Key Applications
CounterQuill
Context: Educational tool for everyday users to write counterspeech against online hate speech
Implementation: Utilizes a three-stage workflow: learning session, brainstorming session, and co-writing session, incorporating AI assistance.
Outcomes: Participants reported increased confidence in crafting counterspeech and a better understanding of hate speech dynamics.
Challenges: Participants faced initial uncertainties and emotional burdens when engaging with hate speech and counterspeech writing.
Implementation Barriers
User-related barrier
Users often feel overwhelmed by the complexity of identifying hate speech and crafting appropriate responses.
Proposed Solutions: The CounterQuill system breaks down the writing process into manageable steps, enhancing user confidence through structured learning.
Technical barrier
Many users, especially those without technical backgrounds, struggle with traditional AI tools that require prompt engineering or manual adjustment.
Proposed Solutions: CounterQuill's intuitive natural language interface minimizes the need for technical expertise, making the tool accessible to all users.
Project Team
Xiaohan Ding
Researcher
Kaike Ping
Researcher
Uma Sushmitha Gunturi
Researcher
Buse Carik
Researcher
Sophia Stil
Researcher
Lance T Wilhelm
Researcher
Taufiq Daryanto
Researcher
James Hawdon
Researcher
Sang Won Lee
Researcher
Eugenia H Rho
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Xiaohan Ding, Kaike Ping, Uma Sushmitha Gunturi, Buse Carik, Sophia Stil, Lance T Wilhelm, Taufiq Daryanto, James Hawdon, Sang Won Lee, Eugenia H Rho
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai