Audiences respond strongly to authenticity on social media. Real photos of your team, products and customers build trust in a way AI images cannot. AI images work best where authenticity is not the point: illustrations, abstract backgrounds, concept visuals and quick mock-ups. The Academy's glossary explains copyright.
Good uses and poor uses
| Good uses | Poor uses |
|---|---|
| Illustrations for educational carousels | Fake product photos |
| Backgrounds and textures | Fake customers or testimonials |
| Concept art for planning | Realistic images of real people |
| Visual metaphors for abstract topics | Before and after results |
| Mock-ups for internal review | Food or venue images that misrepresent |
Common flaws to check
- Distorted hands, faces and text.
- Inconsistent lighting or perspective.
- Garbled logos or lettering.
- Unrealistic product details.
- Overly polished, generic look.
Always zoom in and check before posting.
Prompting for images
Describe subject, style, composition, colours, mood and format. Specify aspect ratio for the platform; Instagram Reels and Stories use 9:16 vertical, 1080×1920✓, while feed posts suit square or portrait. Mention what to avoid, such as text in the image. Iterate on the best results.
Keeping a consistent brand look
Use the same style descriptions, colour references and composition rules every time. Some tools allow style references or saved presets. Combine AI elements with your brand fonts and templates in a design tool. The Academy's guide to a consistent brand look on social media explains visual consistency, and the guide to Canva for social media covers templates.
Copyright and licensing
Check each tool's terms on commercial use and ownership of outputs. Avoid prompts that imitate specific living artists or recreate trademarked characters and brands. Ownership of AI outputs is a developing legal area, so treat AI images as lower-certainty assets and avoid them for key brand elements such as logos. The Academy's guide to copyright on social media covers wider issues.
Labelling AI images
Platforms increasingly label AI-generated or AI-edited content, and some require creators to disclose realistic synthetic media. Where an image could mislead people into thinking something real happened, disclose it. The Academy's guide to AI labels on social media explains platform rules.
Advertising rules
AI images in ads must not mislead. Showing a product that looks better than reality, or results that cannot be achieved, breaches UK advertising rules regardless of how the image was made. The Academy's glossary explains the CAP Code.
Editing real photos with AI
AI editing tools can remove backgrounds, extend images, fix lighting and remove distractions from real photos. This is often lower risk than fully generated images, as long as edits do not misrepresent the product or results.
When to use real photos instead
Use real photography for products, team, premises, customers and events. These images build trust and avoid misrepresentation. The Academy's guide to product photography with a phone shows how to get good results without a studio.
Accessibility
Write alt text for AI images just as for photos, describing what the image shows and why it matters. The Academy's guide on how to write alt text explains good practice.
Image sizes and formats
Generate images at the size and ratio you need, or larger, rather than stretching small outputs. Pinterest favours tall images, LinkedIn and feeds suit square or portrait and Stories need vertical. Check how text overlays look on a phone screen before posting, because details that look fine on a desktop can disappear on mobile.
Text in images
AI tools still struggle with readable text inside images. Generate the image without text, then add headlines and captions in a design tool where you control spelling and fonts.
Combining AI and real assets
Some of the most effective uses combine both: a real product photo placed on an AI-generated background, or real team photos framed with AI illustrations. Keep the real element accurate and the AI element clearly supportive.
Building an AI image workflow
Decide which content types can use AI images, set a style guide for prompts, store approved outputs in a shared library with notes on the tool and prompt used and review images before posting. A clear workflow stops AI images creeping into places where real photos would serve the brand better, and keeps a record if questions about rights arise later.
Audience reactions
Some audiences react negatively to AI imagery, especially in creative industries, food and hospitality, where they expect real craft. Watch comments when you first use AI images and be ready to explain or adjust. Transparency usually works better than hoping nobody notices.
A worked example
A financial adviser creates carousels explaining pensions. Stock photos feel generic, and a designer is too expensive for weekly posts. She uses an AI tool to create simple, consistent illustrations in her brand colours for each slide, adds text in a design template and writes alt text. Engagement rises, and she continues to use real photos for posts about her team and clients' events.
Common mistakes
- Posting images with obvious AI flaws.
- Using AI images to depict products or results.
- Imitating specific artists or brands.
- Not disclosing realistic synthetic images.
- Inconsistent styles across posts.
Learning AI visuals
The Academy's AI tools module in the second level covers image generation and editing, and the phone photography module in the first level covers real photography. The free assessment shows where you would start.
