AI image generated by Ideogram V2

You may not need an image of your CMO wearing a crocheted business suit, but it might be fun to try it anyway. AI-generated images are popping up everywhere — from Instagram feeds to email campaigns, and even in presentations where someone “improvised” instead of grabbing a stock photo. Don’t worry if you’re an AI Debbie-downer, we’ll get more into the dark side of AI image generation in a future post.

For full disclosure, at Twenty44, we purposely generate AI images and use them in our marketing, communication, and deliverables. We treat them as real-world experiments, to test the models and establish best practices to help guide others in their AI journeys. I have been a graphic designer/art director for all my adult life, directing photoshoots, licensing stock images and creating my own illustrations, and I use generative AI with a mix of caution and enthusiasm every day. And yes, I still count the fingers on every image.

Many people who started with AI tools for writing are now curious about expanding into visual content — and it’s an exciting next step. If you’ve been curious about using AI generated images for your own work, you’re in the right place.

The good news? It’s easy to generate images with well-structured text prompts — and thanks to better tools and more accessible platforms, getting started is easier than ever. The catch? Just because you can post AI-generated images doesn’t always mean you should — at least, not without thinking through a few key points first.

Whether you’re a marketer, content creator, or just someone tasked with jazzing up the company newsletter, here’s what you need to know before you start using AI-generated visuals in your communications.

1. Transparency Isn’t Optional (Even if It’s Awkward)

Unlike Fight Club, the first rule of using AI images for your brand is not to keep it to yourselves: Tell people when you’re using AI images for your brand.

It’s not about spoiling the magic — it’s about trust. Your customers, followers, and clients deserve to know when an image was generated using AI. It’s a simple gesture that says, “We’re experimenting with new tools, but we respect you enough to be upfront.”

This can be as low-key as adding a caption like, “Image created with AI,” or including a small icon in the corner of the image. The point is to include a signal that you’re being honest and transparent about your use of AI. 

Remember, you’re not fooling anyone — most AI images are easily detected by sharp-eyed viewers. The obvious giveaways include unnatural hands and fingers, distorted text, overly smooth skin, and odd visual artifacts. More subtle clues can include mismatched lighting, inconsistent shadows, awkward limb positioning, and visual elements that seem almost-but-not-quite realistic.

2. Prompts Are the New Creative Brief

If you want an AI tool to generate something useful, your prompt describing what you want is everything.

A prompt is simply the text instructions you type into the tool to tell it what to make. Think of it as giving creative direction to the AI and reducing the guesswork it has to go through to complete the task.

Here’s a basic recipe for beginner prompts:

  • Subject: What’s in the image? (A cozy coffee shop, a futuristic cityscape, a happy dog wearing sunglasses)
  • Style: Realistic? Cartoonish? Minimalist? Artsy?
  • Mood: Professional, playful, dramatic?
  • Colors: Your brand colors, seasonal tones, or anything else that fits.
  • Use case: Is this for Instagram? A pitch deck? An event invite?

Want to practice? Try describing something ridiculous, like “create an image of a squirrel CEO reviewing quarterly acorn profits in a glass office” — you’ll learn fast what works and what doesn’t. In this example you can see it does a decent job, but you can also see that AI is still not great at generating text.

Image generated by AI (Dall-E 3)

3. Copyright Confusion: Who Owns This Image Anyway?

Brace yourself for some legal murkiness: When you generate an AI image, you don’t necessarily “own” it the way you would a photo you took yourself.

AI tools are trained on oceans of existing images — some of which were gathered without clear consent from the original creators. That means the output can be, legally speaking, complicated. If you’re just experimenting for fun, you’re fine. But for client work, product marketing, or anything with legal stakes? Proceed with caution.

Some platforms (like Adobe Firefly, Shutterstock and Getty Images) offer more commercial-friendly terms. Others leave you to fend for yourself. When in doubt, blend AI images with original content you own — or stick to projects where the stakes are low (like a blog header or an internal mood board).

4. Watch Out for Hidden Bias and Accidental Weirdness

AI loves patterns — but some of those patterns can bake in bias or lazy clichés. Depending on how you phrase your prompt, you could unintentionally generate images that lean into stereotypes or just look…off.

Example: Ask for an “image of CEO” and the AI might give you a suspiciously homogenous image (see the cringe-worthy output posted below). That’s on the training data, not you — but it’s your responsibility to catch it before it hits your feed. As with all AI-generated content, you are ultimately responsible for what you post.

Pro tip: Before posting anywhere, ask yourself:

  • Does this image represent my audience and values?
  • Could anyone find it misleading, offensive, or just plain cringey?
  • Would I feel comfortable explaining how this image was made in a client meeting?
Generic CEO image - generated with AI (Dall-E 3)

Image generated by AI (Dall-E 3)

5. Tools to Try (Without Spending a Dime)

Generating AI imagery doesn’t have to blow your budget. In fact, you can practice for free while you’re learning the ropes. Here are a few beginner-friendly tools to get you started.

  • DALL·E: Even a free ChatGPT account will let you generate a couple of images using Dall-E each day, if you’re on any paid account
  • Microsoft CoPilot: If your company already has an active Copilot license, you can generate images from DALL·E in the app.
  • Google Gemini: Access the Imagen 3 AI model and produce a limited number of images daily before you have to upgrade to a paid plan.
  • Meta AI: Meta AI is a quick option to get experience—just describe what’s in your imagination. If you don’t like what you see, tell the AI what to change. Just remember, the whole image will be change slightly as each new generation is unique.
  • Adobe Firefly: If you use Adobe Creative Cloud applications like Photoshop you can generate images directly in the application. 
  • Canva: There’s a built-in AI image generator included in free accounts.

The quickest and smartest way to learn? Play with prompts, review what works, and avoid the pressure of going public before you’re ready.

Once you’ve gained confidence you can move on to one of these more sophisticated platforms. Most offer a free trial, and the monthly rates are reasonable.

6. When AI Images Shine — and When to Skip Them

AI images are fantastic for:

  • Quick-turn social posts.
  • Brainstorming and pitch decks.
  • Internal content like newsletters or presentations.
  • Concept art and visual experiments.

That said, stock imagery might still be the more obvious choice for certain images, especially since accessible libraries offer high-quality photos and illustrations at low cost.

AI images are risky for:

  • Product photos.
  • Anything tied to a legal claim (like health, safety, or financial products).
  • Representations of real people, especially public figures.
  • Logos and official brand identity work (seriously, don’t).

7. Set Your Own Rules (Because Guidelines Beat Guesswork)

Before your team dives into the AI image pool, first refer to your company’s AI usage policies, and use that to set some house rules

Make a decision:

  • When it’s OK to use AI images.
  • When you’ll require a human designer or photographer.
  • How you’ll label AI images for transparency.
  • Who reviews AI images before they go public.
  • What tools you trust (and which ones are off-limits).

Even if you’re a solo creator, writing down your own rules helps you stay intentional — and avoids future “Oops, we didn’t think about that” moments.

Final Thought: Creativity + Responsibility = Smart AI Use

AI images can save time, spark ideas, and stretch your creative muscles — all good things. But how you use them matters just as much as what you make. If you start with transparency, ethics, and a little creative curiosity, you’ll be well ahead of the pack.

So go ahead — generate, experiment, laugh at the weird results, and find what works for you. Just remember to be upfront about where those images came from. Your audience will appreciate the honesty — and you’ll sleep better at night.

Now, I’m off to do a bit of research into the crocheted business suit market… using AI of course.

This post includes assistance from Perplexity, Google Gemini and ChatGPT. Images generated using Dali-E-3 or Ideogram.

Randall Matheson profile picture

Randall J. Matheson

Randy Matheson is an innovation strategist with a 25+ year proven track record of turning ideas into digital products. He specializes in working with Generative AI for content creation and using cutting-edge AI tools to create and interact with virtual audiences. He operates out Hamilton, Ontario where he resides with his partner and two large dogs.

Connect with Randy on Linkedin