Our approach to photography and artificial intelligence
Behind every story we share are people whose lives, dignity, and trust matter deeply to us. We want our images to help you connect with them and understand the work your generosity makes possible. We also want you to understand how those images are made.
Most of the images we publish are original photographs from our partners and Journeys of Compassion. These remain our first choice. In certain circumstances, we also use artificial intelligence to improve photographs, protect someone’s identity, or create an illustration.
Our guiding principle is simple: the way we tell a story should honor the people whose story it is.
Why we sometimes use AI.
Our grassroots partners are busy caring for people. They may have limited camera equipment, little time for photography, or an internet connection that makes sending large files difficult. A photograph that works on a phone screen may not have the resolution needed for a printed catalog.
Sometimes the considerations are more personal. A photograph could identify a survivor of abuse, expose a child’s circumstances, or put a partner serving in a dangerous environment at risk. During a crisis, asking someone to pose for the right photograph may place an unnecessary burden on a person who needs care.
We use AI selectively in four ways:
Grounded in the communities we serve
When creating illustrations, we use relevant photographs from our work to guide details such as the setting, buildings, clothing, landscape, and activities. These references help us represent the communities we know and love with greater care and specificity.
An image informed by real photographs is still an illustration. Reference material does not make a generated scene a record of an actual event.
We take responsibility for the images we choose to publish. Our review considers accuracy, dignity, cultural context, and whether an image could give a misleading impression. We may also seek feedback from local partners when their knowledge would help us assess a depiction.
AI is never used to invent program results, exaggerate someone’s hardship, or present a proposed project as an accomplished one.
Respecting privacy
Protecting people includes considering how reference photographs are handled during editing. We store original photographs in access-restricted systems. Using an external AI service involves a separate transfer of any images submitted to it, and those services have their own terms and data practices.
Our approach is to limit sensitive information shared with these tools and consider whether a less identifying reference could serve the same purpose. We do not assume that changing a person’s appearance in the finished image resolves every privacy concern.
Recognizing AI imagery
We use an image symbol to identify visuals that have been generated or substantially altered using AI, including changes made to protect someone’s identity. The symbol is explained alongside our image-policy notice in publications and on our website.
Routine adjustments to brightness, color, cropping, or sharpness do not necessarily receive this marker. Its purpose is to help you recognize a meaningful difference between an original photograph and imagery that has been created or materially changed.
An approach we will keep learning from
The technology will change, and our practices will continue to develop. Our responsibility to the people we serve—and to those who support this work—remains.
If an image raises a question, we welcome it. We want to tell these stories with warmth, honesty, and care, and we are committed to reviewing concerns and correcting mistakes.