Our third hackathon revolved around hacking Artificial Intelligence systems (but not in the way you might be thinking!). We invited our BIG Thinkers to find and develop scenarios where marginalized groups of people might be left out of already created AI systems.
Team 1: Self Driving Car Hack
Challenge
Navigation and Reaction — Create a scenario for a situation that may not be recognized by the vehicle system.
Solution
Self Driving Cars are becoming a not too distant future, so this team used their time to research ways for present systems to be improved using drone traffic mapping. They also sough out how legislation must be updated when self driving cars become available to the market.
Prototype
Team Self Driving Car Hack’s Prototyping process
Team 2: Human Resource Hack
Challenge
HR and hiring and filters for resumes — create a scenario for a resume that may not be considered or automatically rejected by an AI system.
Solution
This team found that since a lot of people are considered “unemployable” because we use outdated standards. We have a high volume of applications for job posting and some qualified candidates get screened out for irrelevant reasons. Jobs are also not created to be inclusive, accessible, or flexible in this day and age. In order to remedy this they created a series of questions and a persona to create a more responsive application system.
Prototype
Team Human Resource Hack’s Prototyping process
Team 3: Education Hack
Challenge
Learning analytics and personalized learning — Create a scenario of a learner or educator who may not be considered or automatically rejected.
Solution
This team focused specifically on public schools in El Arrayan, creating a persona and scenario based on already developing educational structures. These remote schools are funded by the government that uses as AI system to filter which schools need more funding. The issue with this system is that it does not take into account differing features of its surrounding region that greatly impact each school. To solve this, this team created a new set of criteria for evaluation for the AI to consider including things like number of students with disabilities and a locations access to technology.
Prototype
Team Education Hack’s Prototyping process
Team 4: Health Hack
Challenge
Hotline decision support systems. Create a scenario of a hotline call that may be misnavigated or disconnected.
Solution
Eloisa created a persona of a trans man who, after a failed suicide attempt, called a hotline in order to enter the in-patient psychiatric ward. The scenario developed went into how the persona would be misgendered, transferred into the incorrect ward, and then face conflict between their hormonal treatment and their anti-depressant medication in regards to applying to hormonal therapy or sex-reassignment surgery. Eloisa’s solution was to create a new series of questions for the AI to learn from.
Prototype
Team Health Hack’ Prototyping process
Team 5: Banking Hack
Challenge
Loan and credit applications — Create a scenario for a person who may be automatically rejected by an AI banking system criteria.
Solution
The team for this challenge developed a detailed persona of a man who is going through a divorce and is being greatly affected financially. This persona has ADHD and because of his lack of a vehicle due to the separation cannot access his medication regularly. Due to his increasing stress, his work suffers and he starts to miss credit card payments that start to impact his credit rating. This team decided to remedy scenarios like this by creating an added set of questions for banks to assess creditworthiness.