Results

Our Work is Having an Impact in AI

AI4ALL Removes Barriers for Underrepresented Young People to Get into AI

AI4ALL Programs are Designed to Create Changemakers

AI4ALL’s programs are deeply rooted in research and continually tested and refined. Our curriculum is developed in-house by content experts and educators, reviewed by leading AI researchers and industry experts, and tested at top AI institutions around the world. AI4ALL alumni have access to a lifelong network of support and are already making significant impacts in the field through research and outreach.

Who We Serve

AI4ALL strives to center the voices of underrepresented people in AI and we are improving each year. This data pertains to our 2018 summer programs cohort.

  • 78% Female
  • 22% Male
  • 40% Asian
  • 27% Latinx
  • 13% African American or Black
  • 11% White
  • 1% Native American
  • 1% Two or more races
  • 4% Declined to identify

Program Goals and Impact

  • Developing Technical Skills

    92% of alumni report they know the distinction between different types of machine learning

  • Nurturing Interest in AI

    77% of alumni are interested in
    a career in AI

  • Creating Belonging and Community

    90% of alumni feel they are a part of a community in AI and computer science

  • Supporting Alumni Starting AI Projects

    61% of alumni have started
    AI/CS projects

  • Connecting to Role Models and Mentors

    93% of alumni have female role models in the AI field

  • Now is a Critical Time for AI

    There is a diversity crisis in AI and computer science: a homogenous group of technologists is building AI solutions for our diverse population. The global economic impact of AI is expected to reach $15 trillion by 2030 1, and we’re already seeing AI incorporated into areas like medical diagnosis, self-driving cars, and customer support. At the same time, only 12% of AI researchers globally are women, and the numbers are even more dismal for people of color in advanced STEM roles: and Black and Hispanic men and women make up under 11% of people employed in science and engineering jobs in the US 2. This lack of diversity results in biased AI products that, at best, don’t serve everyone and, at worst, actively harm underrepresented groups. There is evidence that existing societal biases, including sexism, racism, and other forms of discrimination, are being built into AI products. It is crucial to take action now to ensure that everyone has the opportunity to guide the creation of AI as a tool for good.

Barriers to getting into AI

  • Access: Lack of Awareness and Exposure

    • Only 35% of high schools in the United States teach Computer Science 1
    • Only 1 in 4 students (26%) who attend high poverty schools have access to any CS course in their school 2
    • Without curriculum that is engaging, culturally relevant, or aligned with the interests of underrepresented groups, CS will continue to be perceived as solitary, lacking interaction, and lacking connection to societal challenges 3
  • Interest: Lack of Technical Training, Confidence, and Feelings of Belonging

    • Barriers that prevent underrepresented talent from entering this field including a lack of exposure to technical concepts early; few relatable role models and mentors; and lack of peer-to-peer support 1, 2, 3, 4
    • Students question their belonging in new academic settings, especially when there is stigma or stereotype threat 5
    • AI4ALL program elements align with the 10 effective features of successful K-12 STEM intervention programs for underrepresented groups 6, 7
    • Research shows self-efficacy has the largest impact on STEM entrance, intent to major in a STEM field, and persistence within that major 8
  • Persistence: Lack of Professional Networks, Role Models, and Community

    • Underrepresented students lack access to programs and social networks that prepare and connect them with internship and workforce opportunities 1, 2, 3
    • Research shows the lack of women and underrepresented minorities in computing contributes to the limited peer networks, mentors, and role models for students from diverse backgrounds 4, 5, 6
    • Access to direct college admissions support, in addition to peer networks, has been shown to promote retention and persistence of underrepresented students in STEM 7

AI4ALL Alumni are Changemakers

  • Using AI to Improve Surgery

    Amy discovered AI as a student at Stanford AI4ALL in 2015. She spent the next few years deeply researching computer vision, a type of AI, with mentors at Stanford. In 2017, her research that used computer vision to provide feedback to surgeons and improve outcomes for patients won the Best Paper Award among papers about machine learning and health at NeurIPS 2017. Read More.

  • Using AI to Build Community

    A few years ago, Kevin immigrated to the US from Vietnam with a vision of better access to tech classes. When he found that his new high school didn’t offer many, he joined Carnegie Mellon University AI4ALL 2018. After finishing the program, he was inspired to share his new AI knowledge with his high school classmates, so he started his own AI club with the support of an AI4ALL alumni grant. Kevin has already taught 112 of his classmates about AI through his club. Read more.

  • Using AI to Improve the Criminal Justice System

    At Stanford AI4ALL 2016, Bekah was excited to be exposed to a community of people applying AI to social issues. After completing the program, she was able to use her new AI skills in combination with her interest in examining inequities in the criminal justice system at a research internship at MIT Media Lab on the uses of AI in the criminal justice system. Read more.

  • Using AI to Increase Access to Emergency Care

    Inspired by losing a family member due to excessive wait times for ambulance care, Stanford AI4ALL 2017 alum Viansa joined with 2 fellow alumni to do something about it. The three alumni used natural language processing to create a system to help emergency responders rank the urgency of 911 calls. Their system ensures that people with the most urgent needs are tended to first. Read more.

How We Measure Success

AI4ALL rigorously measures student outcomes using pre- and post-program surveys as well as medium to long-term outcome tracking for our students. We measure success by tracking AI4ALL student progress towards these outcomes.

  • Immediate Impact

    • Increased sense of belonging in the field
    • Development of technical skills
    • Increased confidence
  • Turning-Point Impact

    • Enrollment in post-secondary computer science/AI programs
    • Internships in CS/AI
  • Long-Term Impact

    • Entry into AI and related fields
    • Retention and advancement in AI and related fields