Exploring AI-powered pro-social interventions in online discussions

Online platforms have historically (and infamously) suffered from the problem of so-called "toxic" behavior, such as personal attacks, harassment, and general incivility. The HMC AWeSoMe (Analyzing Web and Social Media) Lab is exploring the potential for AI-powered systems to help users proactively prevent their online conversations from escalating into toxicity, by using machine learning and natural language processing to identify conversation threads that are at risk of turning toxic, then presenting the user with an understandable, justified, and transparent intervention encouraging them to think more carefully about their reply.

This research involves a joint technical and social approach. On the technical side, while models already exist that can detect at-risk conversations with reasonable accuracy, these models are not sufficiently transparent or explainable. Students interested in the machine learning / NLP side of the project will have the opportunity to work with cutting edge LLMs to investigate novel approaches to summarizing conversations, as a way of making model decisions more explainable. On the social side, previous AWeSoMe Lab students have developed a web platform for live testing of AI-powered interventions with human subjects. Students interested in the social side of the project will have the opportunity to develop, and possibly launch, a pilot user study leveraging this platform.

In your essay, please specify which aspect of the project (technical or social) appeals to you (it could be both!), and why. 

Name of research group, project, or lab
AWeSoMe Lab
Why join this research group or lab?

Our work straddles the boundary between computer science and social science. If you are a Harvey Mudd student who is itching for an opportunity to jointly make use of your interests in STEM and HSA subjects, this may be a perfect fit for you! Students with interests in sociology, government, or philosophy are especially encouraged to apply. Furthermore, this project has room for contributions both from students who are new to computer science and from students with more extensive experience. First and second year students will get a chance to apply their CS5 and CS35 knowledge while working with our Python-and-JavaScript-based web platform, while also gaining exposure to the underlying machine learning concepts which could spark a future interest in that area. Meanwhile, more senior students with prior AI/ML/NLP experience will have a chance to gain experience working with cutting edge LLMs.

Logistics Information:
Project categories
Computer Science
Artificial Intelligence
Human-Computer Interaction
Machine Learning
Natural Language Processing
Student ranks applicable
First-year
Sophomore
Junior
Senior
Student qualifications

Students should at minimum have taken CS5; CS70 would be the ideal level. For more advanced students seeking to work on the machine learning / NLP aspects, at least one of the following courses is preferred: CS151 Artificial Intelligence, CS158 Machine Learning, CS152 Neural Networks, CS159 Natural Language Processing (Alternatively, if you have done a project or internship that uses concepts from one of these courses, you may mention that in your essay response and that can substitute for the corresponding course).

Time commitment
Summer - Full Time
Compensation
Paid Research
Number of openings
4
Techniques learned

Web development, user study design, data analysis, working with LLMs, conversation analysis

Contact Information:
Mentor
Jpchang@hmc.edu
Assistant Professor
Name of project director or principal investigator
Jonathan Chang
Email address of project director or principal investigator
jpchang@hmc.edu
4 sp. | 0 appl.
Hours per week
Summer - Full Time
Project categories
Machine Learning (+4)
Computer ScienceArtificial IntelligenceHuman-Computer InteractionMachine LearningNatural Language Processing