Chatbot study consent

Welcome to our study: Using an AI Chatbot Intervention to Reduce Toxicity in Online Communities (Purdue IRB#: 2023-306) This is a research study run by researchers at Purdue University and Stanford University. Dr. Jeremy Foote (jdfoote@purdue.edu) is the principal investigator.

Key Information

Please take time to review this information carefully. This is a research study. Your participation in this study is voluntary which means that you may choose not to participate at any time without penalty or loss of benefits to which you are otherwise entitled. You may ask questions to the researchers about the study whenever you would like. Be sure you understand what you will do and any possible risks or benefits. We are conducting this study in order to understand how chatbots might help to moderate online communities. The study will involve two types of data collection: we will collect your conversation with the chatbot. We will also collect data on your public actions on Reddit.

What is the purpose of this study?

If you have been contacted by our chatbot, it is because our algorithm determined that one of your recent comments displayed signs of toxicity. We are planning to have our chatbot engage in conversations with approximately 2,000 users whose comments have been flagged by our algorithm.

What will I do if I choose to be in this study?

If you would like to be in the study, then you simply respond to the chatbot. When you respond, you will randomly be assigned to one of a few chatbots. After each of your responses, the chatbot will write back to you via modmail messages. Your participation in this study is completely voluntary, and you are free to withdraw at any time without penalty. The conversation will be conducted entirely through Reddit direct messages, and you will have the option to end the conversation at any time.

We will not collect any personal information about you or anyone else as part of this study. We will collect the conversations that you have with the chatbot. We will also collect data on your public behavior (posts and comments) to see whether chatting with the chatbot changes whether participants post toxic content in the future.

How long will I be in the study?

You can stop sending messages to the chatbot whenever you would like. The chatbot will stop sending messages to you after it has sent 50 messages.

What are the possible risks or discomforts?

There are no known major risks associated with participating in this study; while we have designed the chatbot to be non-confrontational, it is possible that the chatbot will act in unexpected ways. It is also worth noting that because conversations occur through modmail, moderators of the subreddit will have access to your conversation with the chatbot.

We will share our findings with others, but we will never make your username public as part of the study or associate anything you say with your username. Any published research or data from this study will use either aggregate data or anonymous identifiers. A data breach is always a risk when storing data, but we will take precautions to minimize this risk as described in the confidentiality section.

Are there any potential benefits?

While we do not anticipate any major benefits, you may enjoy your conversation with the chatbot and you may learn new ways of engaging online. There are no anticipated costs to participate in this research.

Will information about me and my participation be kept confidential?

Your anonymity will be strictly protected throughout the study, and the data collected will be used for research purposes only. The data and study results may be published in academic journals, presented at conferences, or shared with other researchers. However, we will not gather any identifying information, and your username and any personal information you share will be removed before any findings or data are shared publicly. The file connecting your username to your conversations and public comments will be kept on a secure server and only the research team will have access. This file will be destroyed within five years. The project's research records may be reviewed by the study sponsor/funding agency, US DHHS Office for Human Research Protections, and by departments at Purdue University responsible for regulatory and research oversight.

What are my rights if I take part in this study?

You do not have to participate in this research project. If you agree to participate, you may withdraw your participation at any time without penalty. To withdraw, simply stop chatting with the chatbot.

If you would like your data removed form the study, contact us and we will remove it, up until the point that the mapping file is destroyed.

To opt out of participation or withdraw your consent please notify a member of the research team. The names and contact information for the investigators are listed below. May I withdraw consent? Yes. You may withdraw consent at any time by contacting Jeremy Foote (jdfoote@purdue.edu).

Who can I contact if I have questions about the study?

If you have any questions or concerns about this study, please do not hesitate to contact Dr. Jeremy Foote (jdfoote@purdue.edu) at Purdue University.

To report anonymously via Purdue’s Hotline see www.purdue.edu/hotline

If you have questions about your rights while taking part in the study or have concerns about the treatment of research participants, please call the Human Research Protection Program at (765) 494- 5942, email (irb@purdue.edu) or write to:

   Human Research Protection Program - Purdue University 
   Ernest C. Young Hall, Room 1032
   155 S. Grant St.
   West Lafayette, IN 47907-2114