Skip to content
TECHNOARTH Technology in practice.
AI

How to create and edit images with AI (plus prompts for professional photos)

Here’s how to generate and edit images with ChatGPT, Gemini and Copilot: how to describe the scene, improve a photo without changing the face, create a professional profile picture, and the precautions around image rights and deepfakes.

ChatGPT, Gemini and Copilot create images from text and also edit photos you upload: swapping the background, improving the lighting, changing the style. The quality depends a lot on how you describe what you want.

In this article
  1. How to describe an image
  2. A professional profile picture from a selfie
  3. Improving a photo without changing the face
  4. Where to do it
  5. Important precautions
  6. FAQ

How to describe an image

Include, in this order:

  1. Subject: who or what appears.
  2. Setting: where it is, time of day, weather.
  3. Style: realistic photo, illustration, watercolor, 3D, technical drawing.
  4. Framing and light: close-up, full body, aerial view; natural light, studio, sunset.
  5. Format: square (post), vertical (stories), horizontal (banner).
Prompt
Realistic photo of a small computer store on a neighborhood street, late afternoon, golden light coming through the shop window.
A smiling technician helping a customer at the counter. Medium shot, advertising photography style, horizontal 16:9 format.
No text or logos in the image.

A professional profile picture from a selfie

Upload a well-lit, front-facing photo of yourself and ask:

Prompt
Turn this photo into a professional LinkedIn portrait: blurred light-gray neutral background, soft studio lighting, dark business attire.
Keep my face, features and skin tone exactly the same. Don't change my age, face shape or expression.

Improving a photo without changing the face

Prompt
Improve the quality of this photo: increase sharpness, fix the lighting and colors, and remove noise.
Don't change faces, expressions, clothing or the background. Keep the original framing.

Where to do it

  • ChatGPT: generate or upload the image and describe the change.
  • Gemini: in the app or on the website, upload the photo and ask for the edit.
  • Copilot: in the app or at copilot.microsoft.com, with your Microsoft account.

They all have usage limits on the free version. To choose a tool, see ChatGPT, Gemini, Copilot or Claude.

Important precautions

  • Don’t use other people’s faces or voices without permission. Fakes involving real people (deepfakes) can be a crime and violate image rights.
  • AI images in company ads and posts: avoid third-party logos, brands and characters. Make it clear when an image was AI-generated if it could mislead customers.
  • Scams: criminals use fake images and videos of well-known people to ask for money. See how to spot phishing.
  • Personal photos: avoid uploading photos of children and documents to AI tools.

FAQ

Can I use the images AI creates commercially?

It depends on each service’s terms and on the content of the image. Read the tool’s terms of use and avoid images that imitate brands, works or real people.

Why does AI get text inside the image wrong?

The models have improved, but they still often get letters and words wrong. For banners with text, generate the image without text and add the text in an editor (Canva, PowerPoint).

Share

About the author

TECHNOARTH Editorial Team

The TECHNOARTH newsroom: technology tutorials, guides, comparisons and news produced with AI assistance and checked against manufacturers’ official sources.