Social Justice Chatbot Hack

Inspiration

Philadelphia youth hackers are invited to participate to gain a deeper understanding of racial justice. Furthermore, Chatbot Hack provides students with a business model that can be utilized to develop chatbot software services for local businesses in Philadelphia. Participation in the project will increase youths' understanding of their place in the world and how to change it for the better,

What it does

Philadelphia youth can benefit from computer science education and training through the project. The goal of the project is to facilitate the participation of computer science by Philadelphia youth. Mentor will be recruited to join the PASTEM.ORG virtual community. These mentors will reach out to youth to educate, promote and highlight the impact of this program. These trained teacher mentors can offer both training and social bonding for Philadelphia youth.

How we built it

CHATBOT HACK is hosted on SAP Conversational AI Low Code Chatbot for free to users. The virtual assistant can be quickly programmed, deployed, and managed after registering and verifying the student's Gmail address. Philadelphia youth will learn how to create a personalized virtual assistant by using dialogue skills. In SAP Conversational AI, intents are created within the dialogue skill to understand what the student is trying to convey.

In conversational systems, Natural Language Understanding (NLU) involves intent detection and slot filling. The two tasks are used to obtain a structured representation of the meaning of the utterance so that it can be processed by a computer. SAP Conversational AI Low Code Chatbot needs only be given a few examples of how a user might express each intent. Even if the user's utterance does not exactly match what they intended it can recognize their intent once trained.

To have SAP Conversational AI Low Code Chatbot assistant respond, students can add prebuilt intents from SAP Conversational AI assistant. It is easy to add intents to the anything else note by creating a dialog note above it. Once the student has added the virtual assistant's response, she or he is ready to test the virtual assistant. When the same response appears more than once in a conversation, students can add variations to the CHATBOT HACK. Having developed a system that contains the skill, it is simple for the student to make it public.

Challenges we ran into

One of the biggest challenges is finding experts to help with this idea. Through our networks, we were able to receive guidance from SAP AI experts and computer science educators regarding functionality and community engagement. Our team managed to navigate through the IBM Watson Supercomputer site, but due to time constraints was unable to enter conversational language into Watson for analysis. We recieved the following message before being unable to continue:

“Text Analysis with Watson Natural Language Processing Description: In this project, you will find examples on how to use Watson Natural Language Processing models to extract insights from text. You'll also see how you can use Watson Natural Language Processing to create your own models. The project contains one data set as CSV, and five notebooks. 5 notebooks use the latest Python and Watson NLP XS (beta) environment. This environment contains the watson_nlp Python library, and provides access to pre-trained NLP models. One notebook uses a custom environment, based on the latest Python and Watson NLP, to have more CPU and Memory resources available for the notebook.”

Accomplishments that we're proud of

PASTEM.ORG was able to gather a team of STEM education professionals and Drexel CS students. Working together we edited the http://www.pastem.org web site to include a page on AI Literacy: Philadelphia youth can benefit from computer science education and training through the CHATBOTHACK project.

What we learned

Analyzing social media trolling can help Philadelphia youth understand messaging on media, including social media. We use the Generative Pre-Trained Models (GPT 3) natural language model that is built into the SAP Conversational AI platform. GPT 3 can write new text, translate text, chat with users, and even with the proper programming to answer abstract questions. Computer savvy Philadelphia youth with little or no instruction can utilize natural language understanding of the GPT 3 Models to comprehend messages that they are bombarded with by the media.

PASTEM.ORG uses SAP Conversational AI platform’s GPT 3 Modeling for conversational analysis. Using artificial intelligence to understand conversational speech begins with mining the web for information or discovering knowledge from hypertext. Web crawling and indexing are techniques for producing knowledge from hypertext data. Machine learning can be used to systematically understand data acquired and stored. Using SAP Conversational AI Low Code Chatbot, high-speed supercomputers can help users analyze social situations and conversations. These systems that attempt to understand the contents of a document, like a news release, are more sophisticated and exhibit greater complexity. Specifically, conversational AI can identify negative and positive emotional content in messaging, as well as keywords that are used to determine whether a social media message is a troll (negative message) or an emoticon (positive message). The SAP Conversational AI Chatbot can be specifically trained to detect two things. First, if given a conversation the chatbot can detect whether its positive (emote) or negative (trolling). The chatbot determines the emotional content and keywords so the user knows what exactly made it negative. When asked the model can decide whether the speech entered into the bot is positive or negative and identifies the keyword that caused it to have a negative connotation.

Will Smith's "Slap" was seen by millions at the Oscars last week, but its social significance is obscure.During the Philly CodeFest, the PASTEM.ORG team attempted to build a CHATBOT HACK to demystif Will Smith's Oscar slap using AI." Natural Linguistic Understanding (NLU). NLU can help with gaining insights into a problem whether it's developing a shopping list for the grocery store or understanding the impact of racism in our daily lives. Wikipedia defines NLU as, “Natural-language understanding or natural-language interpretation is a subtopic of natural-language processing in artificial intelligence that deals with machine reading comprehension. Natural-language understanding is considered an AI-hard problem

What's next for Social Justice

We will recruit students through science career events and community meetings prior to the start of CHATBOT HACK 2.0. The CHATBOT HACK 2.0 project will ask Philadelphia eSports teams and STEM education programs to suggest students who have shown interest and/or aptitude in science, mathematics, and technology. During the start-up period, there will be team-building activities and information sessions for parents and students. CHATBOT HACK 2.0 information will be posted on social media and distributed at networking events to spread awareness about the CHATBOT HACK 2.0 initiative and provide referrals to the PASTEM.ORG CHATBOT 2.0 HACK internship. The recruitment materials and procedures will emphasize that CHATBOT HACK 2.0 is open to all students, especially black and brown youth of both genders who are underrepresented in STEM disciplines.

The purpose of CHATBOT HACK is to facilitate the participation of computer science teachers in CHATBOT HACK 2.0 by providing remote mentoring as part of the PASTEM.ORG virtual community. The mentors will reach out to CHATBOT HACK 2.0 users to educate, promote, and highlight the program's impact. Mentoring by these trained teachers can help Philadelphia youth build bonds and develop their skills. In the post-pandemic era, addressing socio-technical challenges while supporting the development of youth leaders requires building a digital platform that includes community stakeholders.

Built With

• ai
• conversational
• ibm-watson
• javascript
• sap
• video