From Theory to Practice: How to Address Algorithmic Bias
Experience algorithmic bias firsthand: A unique, immersive learning workshop. This workshop is limited to 35 participants. Click the register button below to join the waitlist.
AI bias arises from human decisions and the data used to train algorithms. This data, reflecting human biases and subjective views, is not neutral. As a result, AI can inherit these biases.
For instance, facial recognition software trained mostly on white faces performs better on white faces. Similarly, the Amazon hiring algorithm perpetuated gender biases due to biased training data. Even efforts to eliminate explicit bias can fail because proxy variables may still result in discriminatory outcomes.
While achieving completely bias-free AI may be unrealistic, we can focus on identifying and addressing known biases and remaining vigilant about hidden ones. This is where Alix Rübsaam expansive and immersive AI Bias workshop plays a crucial role.
In this workshop, participants will:
About the facilitator
Alix Rübsaam
Alix Rübsaam is a visionary in the field of AI and bias mitigation. With a background in creating immersive learning experiences, she co-created the AI Bias Workshop with Ty Henkaline. Her dedication to developing inclusive technology and educating others on the implications of bias in algorithms has made a significant impact in the industry. Currently, she's the VP of Research, Expertise & Knowledge at Singularity University.
Past Webinars
%20(4).avif)
.avif)
%20(2).avif)
%20(3).avif)
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)
From Theory to Practice: How to Address Algorithmic Bias
AI bias arises from human decisions and the data used to train algorithms. This data, reflecting human biases and subjective views, is not neutral. As a result, AI can inherit these biases.
For instance, facial recognition software trained mostly on white faces performs better on white faces. Similarly, the Amazon hiring algorithm perpetuated gender biases due to biased training data. Even efforts to eliminate explicit bias can fail because proxy variables may still result in discriminatory outcomes.
While achieving completely bias-free AI may be unrealistic, we can focus on identifying and addressing known biases and remaining vigilant about hidden ones. This is where Alix Rübsaam expansive and immersive AI Bias workshop plays a crucial role.
In this workshop, attendees engaged in a variety of activities designed to enhance their understanding and skills in the following areas:
- Identifying and Analyzing Bias: Examine how design choices influence algorithmic outcomes and explore the mechanisms that lead to unintended AI consequences.
- Simulation Experience: Participate in real-world technology simulations to make informed decisions about training and designing algorithms.
- Advocating for Responsible AI: Develop socially responsible algorithms and identify strategies to promote ethical AI usage.
- Familiarization: Acquire hands-on experience with the design, implementation, and operation of automated decision-making algorithms and machine learning systems.
Main takeaways:
1. Decision making algorithms are very human
And all that human messiness becomes automated and augmented
2. Bias-free algorithms do not exist
Algorithms operate on biases, but not all of them are intended, desired, ethical, and/or legal
3. To effectively use decision making algorithms, awareness is key
Awareness about your blind spots AND about what your goals are
Resources: Click here for the list of resources from this workshop
Read related: Crafting Ethical AI: Insights from our workshop on Algorithmic Bias
Hosted by
Panelists
