Disclaimer: This project idea is the intellectual property of the CivicAI Team.
Team: CivicAI (CUNY-IBM Watson Case Competition 2018)
Team Members: Connor Rowland, Samuel Bouiss, Kaizhen Tan, Yingqiu Lin, and Yinheng Li from Baruch College
Link to the Business Case Analysis.
Click here to ‘watch this team presenting at the Final Presentation of the CUNY-IBM Watson Competition 2018’.
Project Overview – While the algorithms used in the NYC court system are heavily biased and flawed, the government has realized this and is working to make them better. This past January, Mayor De Blasio signed Bill 1696-A to address bias in algorithms used by the New York Police Department, the city’s courts, and dozens of city agencies. Our group aims to cooperate with the city in improving the bailing algorithm used in courts to meet the bill’s expectations. We plan to integrate IBM Watson into the algorithm, which will produce a more efficient and intelligent way in bail determination.
Problem – The NYC Bail determination system is still flawed and heavily biased. Consequently, many innocent people get stuck in jail for unnecessarily long periods of time, which costs the city a lot of money. Much of this bias is caused by poor algorithm design, as the attributes the algorithms evaluate are biased and insufficient to cover most situations.
Approach to address the problem – There are many benefits to working with IBM Watson to improve the algorithms for determining flight risk. First, it creates a partnership which will ensure fairness and visibility of the algorithms used. IBM Watson is the right choice for this project because they have superior text processing capabilities which will create a greater understanding of the cases being processed. This is the right choice for IBM Watson because this project is incredibly scalable. Given the success in bail, there are other potential areas of the legal system which would benefit from the use of AI and other legal systems who would want to employ the systems. There is even potential for a commercial product for Law firms. Ultimately CivicAI ensures fairness in the legal system and provides IBM with a chance to make a product that could change how courts operate.
Benefits and Impacts – According to the Independent Budget Office, it costs $325 a day to house a single detainee. The City spent $2.6 billion in FY 2017 to house an average of 9,500 people on a daily basis. This motivates our desire to improve the current bail system using CivicAI to promote a cost-efficient way to improve fairness and safeness by reasonably determining flight risk and categorize the defendants. Although decreasing the number of people awaiting trial can’t directly reduce the cost due to the existence of fixed cost, it can still play an active role in reducing the cost to a certain percentage.