AI image-generation tools can create impressive learning visuals in seconds. But without thoughtful prompting, they tend to replicate and even amplify the same narrow visual defaults we’ve been trying to move past – white, male, young, slim, and able-bodied.

That’s not just a technical limitation. It’s a problem of representation – and it has real implications for how people connect with your training content.

If you’re using AI imagery in your L&D programs, it’s time to take a closer look at how your prompts shape the outcomes. The good news? With just a few intentional adjustments, you can help your AI tools generate images that better reflect the diversity of your learners and the world around them.

 

Why AI imagery isn’t automatically inclusive

AI image generators learn by example. Most are trained on vast amounts of publicly available online content – images, captions, metadata. But online representation is already skewed, overrepresenting some groups while marginalizing or stereotyping others. That imbalance becomes baked into the models.

As a result, when you type a vague prompt like “CEO in a boardroom,” you’re unlikely to see a woman of color or someone using a mobility aid. Instead, you’ll often get a parade of suits, sameness, and stereotypes.

These visual clichés aren’t harmless. They reinforce exclusionary norms, signal who “belongs,” and overlook the richness of real-world diversity.

 

Your prompts can shift the outcome: tips for inclusive prompt writing

The solution lies in being intentional with how you prompt. Inclusive prompt writing is the practice of describing identities, roles, contexts, and characteristics that move beyond the AI’s defaults.

Instead of letting the algorithm fill in the blanks, you guide it – thoughtfully, clearly, and inclusively. Let’s explore how.

 

1. Start with clear, descriptive prompts

Before diving into prompting for identity, make sure your basic prompt is detailed and purposeful. Include:

• Visual style: What aesthetic do you prefer? Add descriptors like “realistic photo,” “digital illustration,” or “watercolor portrait” for clarity.
• Tone and mood: Should the image feel formal, upbeat, warm, or dynamic? Align the emotional tone with the message or audience you’re targeting.
• Setting and activity: Where is the person or group? What are they doing? Specific scenarios create more meaningful, grounded visuals.

 

2. Include identity markers thoughtfully

The most effective way to interrupt biased defaults is to describe the people you want to see. You don’t need to list every characteristic, but the more intentional you are, the more representative your results will be.

Consider adding:

• Age: Prompt for a mix of age groups or show older adults in leadership and tech-savvy roles.
• Race and ethnicity: Be specific, or use terms like “multiracial group” or “South Asian woman” to guide diversity.
• Gender: Don’t default to men or women only – include non-binary individuals and counter stereotypes (e.g., male nurse, non-binary engineer).
• Disability: Include both visible and invisible disabilities. For example, a person using a prosthetic limb giving a presentation, or someone with a guide dog commuting independently.
• Body type: Ask for a range of body sizes, especially in empowered or everyday roles like yoga instruction, speaking engagements, or leadership meetings.

🔍Example prompt shift: For example, instead of prompting AI for an image of “people working in a busy office,” you could ask for “a multiracial group of professionals in a collaborative office space, including a larger-bodied woman in her 50s and a man using a wheelchair.”

 

3. Use people-first, respectful language

Language matters – both in how you talk about people and in how the AI interprets your prompts.

Avoid outdated or ableist terms like “wheelchair-bound” or “obese,” which can result in insensitive visuals. Use people-first phrasing such as:

• “A person who uses a wheelchair”
• “A larger-bodied individual”
• “A man with a visual impairment using a cane”

These shifts often lead to more dignified, respectful depictions – and they help reinforce inclusive communication throughout your content.

 

4. Add emotional and cultural context

People don’t exist in a vacuum. One of the best ways to break AI image stereotypes is to include emotional or cultural context.

Instead of just describing what someone looks like, also describe what they’re doing and feeling:

• “A joyful Black father playing with his kids at the park”
• “A confident Muslim woman presenting in a boardroom full of attentive colleagues”
• “A focused young man with Down syndrome preparing ingredients in a busy professional kitchen”

Adding these real-life details creates nuance and vibrancy in the generated images, helping you portray people as active participants in relatable scenarios – not just placeholders.

 

5. Prompt for overall diversity

Even if you’re not focused on one specific identity, you can still guide the generator to produce diverse visuals.

Use phrases like:

• “Multicultural team”
• “Group representing a range of ages, ethnicities, and abilities”
• “Inclusive workplace scene with varied gender expression and body types”

This is especially helpful for group imagery – training photos, conference panels, team meetings – where you want balance and visibility for a broader audience.

🔍 Example prompt shift: You could replace the AI image prompt “panel of experts at a business conference” with “a gender-diverse, multicultural, and age-diverse panel of experts at a professional conference.”

 

Representation isn’t optional – it’s essential for learning

Inclusion is not just a checkbox – it’s a core element of impactful learning and communication. When we overlook representation in visuals, we risk alienating learners, customers, and colleagues. When we get it right, we build belonging, boost engagement, and reflect the people we serve.

AI doesn’t “know” how to be inclusive on its own. But you do. With thoughtful prompting, you can use these powerful tools to create imagery that tells richer, more representative stories.

Want to learn additional ways to make your training more inclusive? Check out these four tips for inclusive online learning.

 

Kat Giroux

Senior Content Writer at Articulate