Content

This section explains how content can contribute to racial bias. If you have learning content in your app, there is a very important body of research on

cultural responsiveness⁠

with which someone on your team should be familiar. Ask around to find out who that is and bring them into the conversation. If you can’t find someone, it is likely that your content, and the way students interact with it, introduces bias into your product.

Unfortunately, there is no quick fix to ensure your content is culturally responsive, and the solutions will differ depending on who makes up your target population. For example, if your product is used in schools with a high percentage of English language learners, it is imperative that your product is relevant to these students’ cultures and perspective. Similarly, schools with a high percentage of low-income students should monitor for references to concepts like “riding in an airplane” or “making an investment” that may not be as familiar to students of lower socioeconomic backgrounds. It is not sufficient to ensure you have a Black History Month or Diwali module in your product. If you distribute learning content in your app, it is essential that someone on your team consider content accessibility to students of all races.

There are an abundance of

resources⁠

on culturally responsive teaching that also apply to content found in EdTech products.

After this section, you'll take a look at your

Datasets⁠

🎯 Goals

Evaluate the cultural relevance of your content

Identify content areas that might lead to racial bias in your data and algorithms

🚧 Caution!

Content doesn’t just mean content students consume but also the content your app uses to evaluate students. For example, if the answers that are considered correct don’t account for dialects (pop instead of soda) or accents (English language learners), students’ answers will be marked wrongーalthough they may have comprehended the material.

Consider cultural responsiveness when developing algorithmically-generated examples. If you use AI to generate examples (using a

GAN⁠

to generate images, for example), it’s critical that you make sure cultural relevance is a criteria for evaluation. Following the same guidelines as for human-generated curriculum will help ensure that AI-generated content is held to the same standards.

✅ Activities for Content

Activity 1: Test Your Content

Test Your Content!

Both qualitative and quantitative assessment is useful for assessing content. You should test content with a diverse group of students and collect feedback on their engagement and comprehension. Look for differences in the way students of different races experience your content.

If you develop your own content, make a list of the diversities you considered in creating your content. Did you consider accents, dialectic vocabulary, cultural norms, and skin tones? Consult the broader field of culturally responsive teaching as a critical part of your creation process. If you source your content from a partner, make sure you ask them these questions. Talk to schools you work with or hope to work with to understand what challenges their students face and investigate whether your content would exacerbate these issues. Teachers already deal with this challenge daily.

Quantitative assessment probably only makes sense for you if you generate your own content (and lots of it). If you do, you should prioritize metrics for your content that reflect these attributes. For example, tag content based on the races present in visual content, the accents in auditory content, the cultures referenced, and geographical contexts that are included. Of course, the more detailed, the better. By logging this data, you can easily see which types of content are lacking, and ensure your content stays inclusive over time. Using this technique, you could quickly find that only 10% of books have Black characters (as is often the case in literacy apps), or that 65% of math activities in your app reference investment scenarios that may be unfamiliar to low income students.

🎯Conclusion: Relevance and Responsiveness

There are many tactics you can take to ensure your content is culturally responsive and relevant for all students. Content can introduce various forms of bias into your product experience and to your algorithms if they are more accessible to some students compared to others. Many schools now provide training to help teachers evaluate the cultural relevance of their content and have equity teams dedicated to the evaluation of curricular materials for this very purpose. Rigorously evaluating your own content will ensure your product meets the needs of districts and their students.

Now that we've evaluated our ideas and content, let's look at your
`Datasets`⁠
.

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.