Joy Buolamwini wins national contest for her work fighting bias in machine learning

2017-01-17 · Source: MIT News - Object recognition · Field: Technology & Digital — Artificial Intelligence & Machine Learning, AI Ethics & Social Impact · Depth: Fundamental Awareness, medium

Summary

MIT Media Lab graduate student Joy Buolamwini won a grand prize in The Search for Hidden Figures national contest, receiving a $50,000 scholarship. The contest, inspired by the film "Hidden Figures," recognized Buolamwini's work on fighting bias in machine learning algorithms. Her research addresses the "coded gaze," where facial recognition software often fails to detect features of darker-skinned individuals, leading to exclusionary practices. Buolamwini launched the Algorithmic Justice League (AJL) to highlight such biases, provide a platform for those affected by coded discrimination, and develop accountability practices for AI system design and deployment. She plans to use the scholarship to create "bias busting" tools to combat machine learning bias.

Key takeaway

For AI Ethicists and developers working on computer vision systems, Buolamwini's work underscores the critical need to proactively address algorithmic bias. You should prioritize inclusive training data and implement bias-checking tools during the design, development, and deployment phases to prevent discriminatory outcomes and ensure equitable performance across all user groups. Consider integrating diverse perspectives into your development teams to identify and mitigate "coded gaze" issues.

Key insights

Algorithmic bias, particularly in facial recognition, leads to discriminatory outcomes and requires active intervention.

Principles

• Inclusive participation matters in STEM.
• Storytelling can change cultural perceptions.
• Poets illuminate uncomfortable truths in tech.

Method

Buolamwini's approach involves launching initiatives like the Algorithmic Justice League to highlight bias through media, gather user experiences, and develop accountability practices for coded systems, alongside creating "bias busting" tools.

In practice

• Check system performance across diverse groups.
• Identify and point out cases where bias occurs.
• Develop tools to check for bias in system design.

Topics

• Algorithmic Bias
• Facial Recognition
• Algorithmic Justice League
• AI Ethics
• Machine Learning Fairness

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Ethicist, AI Researcher, AI Student

Related on AIssential

• Bias in the machine (edited) — Bias in AI systems stems from historical human prejudices embedded in training data and model design, leading to real-world harm.
• Statistical and Structural Approaches to Algorithmic Fairness — Algorithmic fairness must move beyond isolated individuals and deterministic audits to address systemic inequalities.
• An Industry Benchmark for Data Fairness: Sony’s Alice Xiang — Ethically sourced, diverse datasets are crucial for effectively measuring and mitigating bias in AI models.
• Essex police pause facial recognition camera use after study finds racial bias — Racial bias in live facial recognition systems can lead to disproportionate identification of specific ethnic groups.
• AI doesn’t create bias, it inherits it – how do we ensure fairness when it comes to automated decisions? — AI fairness is a complex, context-dependent, and evolving challenge, not a purely technical problem.
• Promise and Perils of Using AI for Hiring: Guard Against Data Bias — AI in hiring offers efficiency but demands careful implementation to prevent discrimination.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT News - Object recognition.