Back

How to Edit and Transform Photos with ChatGPT

Practical AI Tutorial for M&E and International Development

How to Edit and Transform Photos with ChatGPT

Learn how to turn field, programme, training, workshop and vacation photos into recognisable illustrations while protecting accuracy, dignity, privacy and the integrity of development evidence.

Quick answer

Upload a photo to ChatGPT, describe the visual transformation, identify what must remain unchanged, review the result carefully and request focused corrections. For programme or humanitarian images, treat the result as a communication product rather than primary evidence, and disclose substantial AI-assisted changes.

What you will learn

Upload safely
Choose appropriate photos and check consent.
Write a strong prompt
Describe the style, purpose and protected details.
Review the output
Check people, objects, setting and meaning.
Use images responsibly
Avoid fabrication, stereotypes and privacy risks.

The six-step workflow

1. Upload
Add the original photograph
2. Describe
State the intended transformation
3. Protect
Name the details that must remain unchanged
4. Generate
Create the first illustrated version
5. Review
Compare the result with the original
6. Refine and disclose
Correct inaccuracies and label substantial changes

Important distinction: communication image or evidence?

An AI-transformed illustration may support a report, presentation, training manual or social media post. It should not replace the original photograph in an evidence archive, monitoring record, verification file or audit trail.

Keep both versions: retain the original image, consent information, date, location, editing notes and approved final version.

1

Upload an appropriate photo

Start with a clear image that you have permission to use.

Open ChatGPT, use the attachment option and upload the photograph you want to edit. You may use a programme photo, community workshop image, monitoring visit, training photo, field activity, conference image or personal vacation photo.

Usually appropriate

  • Training and workshop photos with consent
  • Public events and programme activities
  • Non-sensitive infrastructure or landscape images
  • Personal photos you own or may edit

High-risk or unsuitable

  • Identifiable children without valid consent
  • Survivors, patients or people facing protection risks
  • Confidential documents, forms or identity details
  • Photos from insecure or politically sensitive locations

Development-sector check: confirm that consent covers the intended use, including public communication and substantial visual transformation where relevant.

2

Describe the transformation

Tell ChatGPT what style you want and how the image will be used.

A useful image-editing prompt explains the desired style, the intended audience, the output format and the important features that must remain recognisable.

Prompt-building diagram

1. Style: editorial, playful, realistic or storybook
2. Purpose: report, training, learning brief or social media
3. Keep: people, setting, activity and important objects
4. Avoid: invented people, equipment, services or results
5. Format: square, portrait or wide
Transform this photograph into a professional editorial illustration for a programme learning brief. Keep the participants, facilitator, seating arrangement, clothing, meeting materials and venue recognisable. Use natural colours and clean outlines. Preserve the original number of people and the condition of the setting. Do not add logos, equipment, services or results that are not visible in the original image.
3

State what must not change

Protect the factual meaning of the original scene.

Image transformations can unintentionally alter people, objects and environmental conditions. Reduce this risk by explicitly listing the details that must remain faithful to the source photograph.

Number of participants
Why: avoids exaggerating reach or attendance.
Prompt: “Preserve the original number of people.”
Infrastructure condition
Why: prevents an unfinished facility from appearing complete.
Prompt: “Do not make the building newer or more complete.”
Assistive devices
Why: protects accurate representation of disability inclusion.
Prompt: “Keep the wheelchair and accessibility features visible.”
Cultural details
Why: reduces stereotyping and identity distortion.
Prompt: “Preserve clothing, skin tones and cultural details.”
Activity and materials
Why: keeps the programme activity accurate.
Prompt: “Do not add supplies, tablets or services.”
4

Generate the first version

Submit the prompt and inspect the overall direction.

Ask ChatGPT to create the transformation. When the output appears, first check the general style, composition and recognisability. Do not assume that a visually attractive result is factually accurate.

For a localised change, identify the area precisely in your instruction. For example:

Remove only the distracting vehicle in the far-right background. Keep the people, road, buildings, vegetation and lighting unchanged.
5

Compare the output with the original

Use a structured human review before publication.

Human review diagram

Original photograph
Primary source record
AI-assisted version
Communication output
Human review
Accuracy, dignity, privacy and protection
Approve or revise
Record the decision and disclosure

Check whether ChatGPT changed any of the following:

Number and position of people
Faces, age or gender presentation
Skin tone, clothing and cultural details
Disability and accessibility features
Logos, labels and written information
Equipment, supplies and services
Condition of buildings and infrastructure
Emotional tone and apparent level of success
6

Refine, approve and disclose

Request focused corrections instead of restarting with a vague prompt.

Use short follow-up instructions that address one or two problems at a time:

“Restore the original number of participants.”
“Keep the faces closer to the source photo.”
“Reduce the cartoon effect and use restrained colours.”
“Restore the original condition of the facility.”

Where the transformation is substantial, label the image clearly so viewers do not mistake it for an unedited documentary photograph.

Examples for M&E and international development

Use these prompts as starting points and adapt them to the original photograph, consent conditions and communication purpose.

Community consultation

Useful for learning briefs, stakeholder presentations and participatory evaluation materials.

Transform this community consultation photo into a professional illustration. Keep the facilitator, participants, seating arrangement, venue and discussion materials recognisable. Preserve attendance and do not add equipment or branding.

Enumerator training

Useful for survey manuals, data-quality training and fieldwork presentations.

Turn this enumerator training photo into a clean educational illustration. Keep the trainer, trainees, tablets and room recognisable. Do not add devices or change the number of participants.

Infrastructure monitoring

Useful for monitoring reports, donor updates and inspection guidance.

Create a realistic editorial illustration of this monitoring visit. Keep the field officer, facility and surrounding conditions unchanged. Improve clarity without making the infrastructure newer, cleaner or more complete.

Focus group discussion

Useful for qualitative research training and methodology sections.

Create a respectful illustrated version of this focus group. Preserve the participant count, positions, room layout and discussion activity. Reduce facial detail while keeping participant roles understandable.

Humanitarian distribution

Useful for situation reports, process manuals and staff learning.

Transform this distribution photo into a respectful editorial illustration. Preserve the location, staff roles, queue structure and visible supplies. Do not add beneficiaries, vehicles, supplies or logos.

Agriculture and livelihoods

Useful for extension guides, programme reports and climate adaptation materials.

Convert this field demonstration into a clear development-sector illustration. Keep the farmers, crops, tools and field conditions recognisable. Do not make the crops healthier or more productive than shown.

Public health activity

Useful for awareness materials, training and health programme communication.

Transform this community health activity into a professional illustration. Keep the health workers, participants, equipment and setting recognisable. Do not add medical devices or services.

Education programme

Useful for learning reports, teacher training and school programme materials.

Turn this classroom activity into a friendly educational illustration. Keep the teacher, learners, classroom layout and materials recognisable. Do not add students, technology or resources.

Playful vacation-photo example

You can use the same process for a personal travel or vacation photo. In this case, the transformation may be more creative, but you should still preserve recognisable people, obtain permission from others shown and avoid adding misleading details.

Turn this vacation photo into a playful illustrated scene. Keep the people’s faces, clothing, poses, landscape and main landmarks recognisable. Use bright colours, soft outlines and a friendly storybook style. Preserve the original composition and do not add new people, buildings or objects.

Privacy-preserving transformations

Turning a photograph into an illustration may reduce facial detail, but it does not automatically anonymise the people shown. Identity may still be inferred from clothing, location, assistive devices, body position, name tags or distinctive background details.

Crop
Remove unnecessary identifying areas.
Simplify faces
Reduce identifiable facial characteristics.
Remove identifiers
Hide names, plates and confidential forms.
Generalise location
Reduce distinctive details when necessary.
Transform this photograph into a simplified editorial illustration. Preserve the activity, participant roles and accessibility features, but reduce identifiable facial details and remove visible names, badges and confidential documents.

What to avoid

  • Adding participants who were not present.
  • Making a poorly attended event appear full.
  • Adding equipment, supplies, services or branding that did not exist.
  • Making unfinished infrastructure appear completed.
  • Removing visible problems that are relevant to monitoring or evaluation.
  • Changing demographic, cultural, gender or disability characteristics.
  • Exaggerating emotions, vulnerability, poverty or programme success.
  • Publishing a transformed image as if it were an untouched documentary photograph.

Copy-and-paste master prompt

Replace the text in brackets with details from your photograph and intended use.

Transform this photograph into a [professional editorial illustration / playful educational illustration / realistic digital illustration / storybook illustration / simplified process illustration]. Keep the [people, clothing, positions, activity, objects and location] recognisable. Preserve the original number of participants and the factual condition of the setting. Do not add new people, services, supplies, infrastructure, equipment, branding or results. Remove only [specific distracting element]. Use [preferred colours, mood and visual style]. The image will be used for [evaluation report / programme report / training / presentation / learning brief / campaign / social media]. Maintain a respectful, inclusive and accurate representation of everyone shown. Create the output in a [square / portrait / wide] format. Do not add text unless explicitly requested.

Responsible-use checklist

✓ Appropriate consent covers the intended use.
✓ The activity and setting remain factually accurate.
✓ No people, services or results were invented.
✓ Privacy, security and protection risks were reviewed.
✓ The image avoids stereotypes and preserves dignity.
✓ Accessibility and cultural details remain accurate.
✓ A human reviewer compared both versions.
✓ Substantial AI-assisted changes are disclosed.

Suggested disclosure wording

Light transformation:

This image was adapted from an original programme photograph using AI-assisted illustration tools. The visual style was changed for communication purposes without intentionally altering the activity represented.

Substantial transformation:

AI-assisted illustration based on an original programme photograph. This image is a communication product and should not be treated as primary programme evidence.

Frequently asked questions

Can ChatGPT edit an existing photograph?

Yes. Upload the image and describe the changes you want. You can request a full visual transformation or identify a specific part of the image that should be changed.

How can I keep people recognisable?

Ask ChatGPT to preserve facial characteristics, clothing, poses and positions. Compare the result with the original and request focused corrections when identities or details have changed.

Can an AI-edited image be used as M&E evidence?

A transformed illustration should normally be treated as a communication product, not primary evidence. Retain the original photograph and supporting documentation in the evidence archive.

Does turning a photo into an illustration anonymise participants?

Not necessarily. People may remain identifiable through their clothing, location, body position, assistive devices or surrounding details. Conduct a separate privacy and protection review.

Should AI-assisted images be labelled?

Label substantial transformations when viewers could otherwise mistake the image for an untouched documentary photograph. The disclosure should match the degree of editing and the communication context.

What is the best prompt for a vacation photo?

Specify the illustration style and ask ChatGPT to keep the people, clothing, poses, setting and landmarks recognisable. State that no new people, buildings or objects should be added.

Official product guidance

For current instructions on creating and editing images in ChatGPT, consult OpenAI’s Images in ChatGPT guide.

Continue learning with EvalCommunity Academy

Build practical and responsible AI skills for evaluation, programme management, development and humanitarian work.

Final takeaway: improve the presentation, but do not change the truth of the people, activity, programme, setting or result represented.
Academy

The courses and articles are developed by a team of experienced evaluators, collaborators, authors, and software developers, guided by Fation Luli. EvalCommunity Academy combines practical expertise in Monitoring & Evaluation and International Development with the latest advances in AI to create high-quality, accessible, and practical learning experiences for professionals worldwide.