Police Accountability

Raheem AI

The Situation

Blacks are killed by police at disproportionate and persistently high rates. However, a vast majority of incidents don’t result in the officer being charged with a crime. While some cases emerge into the public view, most instances of police brutality get lost in the dust and grieving families and communities see no repercussions for the perpetrators. There is no meaningful dataset showing how police interact with their precinct aside from the most fatal of incidents, death.

Who We Are

Raheem AI's Facebook Messenger chatbot allows anyone to anonymously rate and report their interactions with the police, whether positive, negative, or neutral. These results are mapped onto their dashboard, which enables both residents to better understand safe locations and routes and garner a better sense of their surrounding police makeup and also police departments and governing bodies to use this data to reform themselves. Raheem holds police and lawmakers accountable for their actions.

The Project

The Raheem chatbot allows users to anonymously rate their interactions with the police. Results from the chatbot are plotted on the Raheem dashboard. However, the chatbot is still in beta and can be updated to become more powerful & support more features. Your project will be to add to the chatbot to make it smarter, respond more efficiently, and give more useful information.

Level 1: NLP

Implement Natural Language Processing (NLP). The current chatbot includes multiple choice replies, which restricts the depth of information and data collected. A successful Level 1 project will be a Raheem chatbot that does not have only multiple choice replies but, instead, will have at least half of the questions answerable by free-form text, parseable by the backend.

Level 2: Machine Learning

Implement Machine Learning (ML). The current chatbot offers a survey with a few elaborate decision-trees. A successful Level 2 project will implement ML to offer a more robust, inclusive conversation for users about their experiences with police.

Level 3: Recommendation Algorithm

Add recommendation algorithm. The recommendation algorithm connects users to opportunities of change, localized to their community. This feature will exist in Raheem's Community Dashboard. By connecting users to local opportunities, you'll help develop new and strengthen existing relationships within communities that increase participation in the fight for police reform. Some suggestions include recommending residents to:
  • write their council member
  • register to vote
  • volunteer for Raheem street team
  • participate in an upcoming protest
  • attend city hearing on police reform
  • join a local organization for police reform
  • apply for local civilian review board
  • Recommended Skils (not all required):
  • Ruby on Rails/Web Development Experience
  • React
  • Experience with Facebook Chatbots (would be nice)