Davos 2019 - Compassion through Computation: Fighting Algorithmic Bias

TL;DR

The algorithmic justice league, led by Joy Buolamwini, and Justine Cassell shed light on algorithmic bias and the lack of diversity in AI technology, emphasizing the need for inclusivity in data and algorithms to prevent discriminatory practices.

Transcript

hello everybody welcome to the session on compassion through computation fighting algorithmic bias I'm Gideon Litchfield I'm the editor of MIT technology review and I'm gonna be moderating the discussion with two very interesting speakers joy Boram weenie who is currently doing she says her fourth degree at MIT and it's founder of the algorithmic j... Read More

Key Insights

• 🥺 Algorithmic bias can lead to exclusionary experiences and discriminatory practices, highlighting the need for more inclusive and ethical AI systems.
• ❓ The representation of diverse voices and values in technology is crucial to prevent biased outcomes and promote innovation.
• 👨‍🔬 Companies' responses to research on algorithmic bias have varied, highlighting the need for increased transparency and accountability.

Questions & Answers

Q: What is the coded gaze and how does it relate to algorithmic bias?

The coded gaze refers to algorithmic bias that emerges from the priorities, preferences, and prejudices of those with power in shaping technology. It perpetuates exclusionary experiences and discriminates against certain groups, leading to biased outcomes.

Q: How accurate are facial analysis systems in detecting gender and skin type?

Facial analysis systems from companies like IBM, Microsoft, and faceplusplus show varying levels of accuracy in detecting gender and skin type. However, all systems perform better on male faces and lighter skin tones, revealing bias against darker-skinned individuals and females.

Q: How have companies responded to the research on algorithmic bias?

Companies have shown different responses to the research. While some have been unresponsive, others, like IBM, have been proactive in addressing biases and improving their systems. However, more transparency and accountability are needed from companies to prevent algorithmic discrimination.

Q: What steps can be taken to address algorithmic bias and increase inclusivity in AI?

Efforts should focus on inclusive data sets, intersectional analysis, and diverse representation in technology development. Legislation can play a role in increasing transparency, ensuring affirmative consent, and preventing the misuse of facial analysis technology.

Summary & Key Takeaways

• Joy Buolamwini, founder of the algorithmic justice league, conducts algorithmic audits to hold companies accountable for algorithmic bias and exclusionary AI systems.

• Buolamwini highlights the coded gaze, which reflects the prejudices and biases of those who shape technology, leading to discriminatory practices.

• Justine Cassell, associate dean at Carnegie Mellon, stresses the importance of technology representing diverse voices and values, as it influences decision-making and societal biases.