AI tools for visual content are evolving quickly. From campaign graphics to social media posts, platforms like Midjourney, ChatGPT, and Pica AI now offer fast and flexible ways to generate creative assets. But while the speed is impressive, AI raises important questions about authenticity, quality, ethics, and creative intent.
I recently completed an AI prompting course with Superhuman to better understand how these tools work. I wanted to understand not just how to use them, but also how to think critically about them. This post presents insights I walked away with and five questions I believe every communicator or creative should ask before using AI-generated visuals.
1. What is this image for, and does AI actually add value?
Before using an AI-generated image, ask: Does this help your message based on its purpose, audience, and impact? Is it appropriate for the people you’re trying to reach? If the image is for a public-facing campaign, does AI bring something unique that stock or original visuals can’t? For example, it brings something unique when your concept is too specific, abstract, emotional, or metaphorical for traditional visual tools. If not, it might not be worth the shortcut.
Example: In June 2025, Grammy-winning artist Erykah Badu and producer The Alchemist released their single “Next To You” with AI-generated cover art, and fans immediately called it “cheap” and “disappointing.” It sparked backlash across social media, with many arguing that the image felt impersonal and undermined the music’s creative value. The lesson: AI-generated visuals can backfire if they don’t align with your brand, audience, or expectations. In this case, it might have worked better as a conceptual draft, rather than the final product.
2. Is this visual authentic to our audience and values?
AI is trained on the internet, which means it often reflects the same biases, stereotypes, and norms baked into mainstream media. This includes things like unrealistic beauty standards, Western-centric defaults, and flattened representations of diversity.
This matters a lot, especially when working on equity-centered projects. Using AI-generated images of diverse people can feel inauthentic or worse, hypocritical, if your goal is to genuinely reflect the communities you serve.
Example: In May 2023, Amnesty International published AI-generated images alongside news of protests in Colombia. The intention was to protect the identities of protesters. However, critics argued that the visuals felt “fake,” with warped faces and inaccurate uniforms, and risked undermining Amnesty’s credibility. Despite labeling them as AI, Amnesty ultimately removed the images and apologized, highlighting how even well-intentioned uses of AI can clash with your values without careful consideration.
Tip: Use AI to supplement, not replace, authentic and inclusive imagery. Always ask: Is this aligned with our mission, or could it do more harm than good?
3. Where did this image originate, and do we have the right to use it?
AI platforms pull from massive datasets – millions of images, many of which aren’t licensed or credited. So even if your output looks original, it may be built from content not approved for public or commercial use.
This raises ethical and legal concerns, especially when working on paid campaigns or branded content.
Example: Disney and Universal recently sued Midjourney, claiming it allowed users to generate AI images that copied protected characters like Darth Vader and WALL‑E. The lawsuit clarifies that AI doesn’t give you a free pass on copyright.
Tip: Always read the tool’s terms (Midjourney, for example, varies licensing by plan), and avoid generating recognizable IP unless you’re licensed.
4. Are we being transparent about what’s real vs. generated?
As AI-generated content becomes increasingly realistic, it becomes harder for audiences to distinguish between real and fake content. In communications, that’s risky, especially when trust and transparency are at stake.
Example: In July 2024, Elon Musk shared an AI-generated deepfake video of Vice President Kamala Harris on X (formerly Twitter), featuring fabricated statements like “I am the ultimate diversity hire.” Though intended as a parody, the video wasn’t clearly labeled as AI-generated and quickly went viral, reaching over 130 million views. The backlash was swift. The controversy even prompted California to pass legislation requiring disclosure of AI-altered political content.
Tip: If your AI-generated visuals could be mistaken for authentic images, label them clearly. Consider stylized illustrations or surreal graphics instead of hyper-realism to reduce confusion.
5. Are we reviewing, refining, and using AI with purpose?
AI can create content quickly, but speed isn’t the same as quality. Never publish the first result an AI tool gives you. Instead, treat it like an early rough draft: refine it, align it with your visual strategy, and ensure it fits your tone and values.
Human intelligence is irreplaceable here, and frankly, I find that reassuring. For now, AI still requires our ideas, direction, and storytelling to truly function effectively. We give it meaning. We give it soul.
Using AI-Generated Visuals Responsibly
AI is powerful, and when used thoughtfully, it can be a real asset, especially for resource-strapped teams like nonprofits and other mission-oriented organizations. But speed and savings aren’t everything. Ask the right questions. Use it with care. And above all, stay thoughtful.