Anti-Racism and Technologies

Let’s look at the role technologists play in contributing to anti-racism.

System Designers and Architects

Designers determine how audiences interact with products and services. System designers and architects, however, determine interactions between different parts of the product or service, and ensure the quality of a product or service. Their roles can include:

• Data Architect
• System Designer
• Solution Architect
• Technical Architect

They will be considering questions such as:

• How will data be stored?
• How much storage do we need?
• How will data be structured?
• Are the developers building the product according to the design?

How System Designers and Architects Contribute to Anti-Racism

People in these roles may be determining datasets, data structures, and how data will be used. To contribute to anti-racism, they can:

• Ensure it is possible for data to be stored in a way that includes attributes that cover a wide range of people (e.g. it may be necessary to include Tribe).

Developers and Other Implementers

Developers and other implementers receive all the information from different teams and build the product or service. They are responsible for a range of tasks, including:

• Implementing algorithms
• Scrutinising the design and architecture
• Designing, developing, and testing solutions
• Fixing bugs and defects in systems
• Ensuring that development is iterative – i.e. sending outputs back to -researchers/designers and putting them in front of focus groups

How Developers Contribute to Anti-Racism

To answer this question, let’s look at the example of machine learning for image processing. In 2015, the Image Labelling function in Google Photos was classifying Black people as gorillas. What role could developers have played to avoid this racist outcome?

Algorithms and Anti-Racism

This occurred because the classification algorithm required modifications to correctly classify dark skin. To avoid this and contribute to anti-racism, developers can:

• think about all types of users when creating algorithms

Datasets and Anti-Racism:

Developers at Google addressed this problem by removing gorillas from the datasets. In this situation, it ‘fixed’ the issue (1). However, this was a short term fix that limited the service rather than tackling the underlying issue.

Image labelling software is complex and difficult to perfect, but to contribute to anti-racism developers can:

• ensure that datasets are wide ranging and that the software is trained and tested by a diverse group of people.

Testers

Testers provide quality assurance, ensuring that a product or service is fit for purpose and is suitable for the audience. Testers typically:

• Ensure that the whole product works end to end
• Confirm that the product adheres to the requirements
• Test products and services with different user groups

How Testers Contribute to Anti-Racism

To support anti-racism, testers can:

• Test with a wide range of user groups to include a variety of views, perspectives and experiences
• Test products with a range of datasets, and ensuring that the datasets that are used are broad enough to ensure the product is suitable for all audiences

How did these roles compare to the people responsible for creating anti-racist technology that you have previously come across? Did anything surprise you? Were there roles that you hadn’t considered?

References:

1. James Vincent 2018. Google ‘fixed’ its racist algorithm by removing gorillas from its image-labelling tech, The Verge.