The Hunt for Hate Speech: Local High School Students Learn from CS Professor

Image

With more time spent in virtual social settings, high school students are more likely than ever to be exposed to extremist propaganda and messaging by way of hate speech. As the issue of online hate speech becomes ever more prevalent, big tech companies such as Twitter and Facebook are investing in artificial intelligence (AI) to curb hate speech on their platforms. But few students in grades 10 – 12 are exposed to exactly how platforms like Instagram, with approximately over 65,000 postings per minute, actually work to combat offensive language leaving students unfamiliar with potential gaps in the algorithms and susceptible to the hate. 

This spring, Roosevelt High School students demonstrated original code created to combat hate speech online. The eleventh graders wrote lines of code that could detect and discourage hateful language in online chat rooms. Students then virtually presented their programming to a  panel of industry experts to identify the importance of thinking critically about the social media they consume and create.

This panel of students creating code at Roosevelt High School is the third of Professor Xiang Ren’s three-stage, three-month long undertaking to help students understand the potential uses and challenges of building an AI system. The project involved a partnership with two local high schools; Roosevelt High School and the STEM Academy of Hollywood, to create a symbiotic relationship between the students and the code.

The former Snapchat data science advisor, Ren decided to work with high school students to showcase  the possibilities of what AI can do.

Benefits to being exposed to STEM in the classroom reaches well beyond test scores and college readiness, engaging in the subject matter has shown to improve problem-solving skills, creativity, mental alertness, and teamwork and collaboration. Unfortunately, not all students are granted exposure to STEM programming through their high school. Only 47% of high schools in the country offer computer science courses and those without are more likely to be located in urban areas populated with non-white students from low-income households. 

What percentage of teens use the Internet all of the time?

45% of teens now say they are online on a near-constant basis. [Pew Research]

Learn More

“We realized how much impact we could make outside of the research community,” says Ren. “Through these different interactive activities, I can inspire [the students] to like the subject matter [hopefully] they will fall in love with the power of artificial intelligence too,” the professor added. 

This connection was a two way street. Hearing from high school students about their experiences and opinions about social media helped Professor Ren make better informed, inclusive decisions in his work. Incorporating marginalized experiences during the early processes of coding makes for a less biased, all-encompassing algorithm. 

“I now realize what the users really need”, says Professor Ren, on the benefit of including the student’s diverse perspectives in his research. Especially important in the ever expansive fight against hate speech, inclusion can be the deciding factor in which communities are protected by the algorithm or not.

Stage one of Professor Ren’s program kicked off in December 2020 with a Zoom presentation about artificial intelligence and hate speech detection. Ren partnered with the USC Viterbi K-12 STEM Center to present to over 100 students at the two local high schools during Computer Science Education Week.

“The presentation overall was really interesting, I never thought about how the way we choose our words can hurt others,” says a junior from STEM Academy Hollywood.

Image
The second stage began in the spring semester, as students started thinking about how to use AI for good. USC student volunteers met virtually with students to serve as mentors and aid in their projects. Volunteers from the National Society of Black Engineers (NSBE) met with students from Roosevelt High School to help troubleshoot code, while Ph.D. students working in Professor Ren’s lab aided STEM Academy Hollywood student groups in drafting and contextualizing their AI for Social Good projects.

Together, USC student computer science volunteers along with the high school teachers were able to support student groups in developing their own projects. 

One volunteer had this to say: “I really love helping kids become interested in CS. So I loved the opportunity to listen to them talk about their interests or cool technology they wanted to learn more about. I also thought it was so cool to see high-schoolers focusing on such an important topic!” 

Computer science teachers from the high school partners noted the engagement of their students in the topic and were determined to keep the learning going. Teachers integrated the presentation into interdisciplinary student projects about computer science, social studies, the arts, and more.

A large part of the benefit, says Mr. Ed Kim, the Computer Science teacher at Roosevelt, are the real world issues brought into the classroom. “Providing authenticity in project topics proved to be important, and it does not get any more authentic than the real-world,” Kim said.  “Listening to USC student testimonials about their journeys in computer science inspired our students to continue challenging themselves,” he added.

USC Viterbi-teacher partnerships were created with LAUSD Linked Learning coordinators Neil Little and Lindsay Corcoran to build the capacity of educators and the community to integrate rigorous student-driven academics, career applied skills, and work-based learning.

Learn More

Professor Ren sent a pre-recorded video to the closing event in which he congratulated students on the completion of their projects and shared the many ways they could use AI for social good in the future.

“Together I would hope that AI technologies help humans to build a better society so we could spend more time with the important people in our lives and the most important parts of our work”, says Ren. 

Published on May 26th, 2021

Last updated on November 12th, 2022