Skip to Content

Transparency and media ethics

Image sources and use of AI

How hpsmedia selects images, documents usage rights and transparently labels AI-assisted content.

Last updated: 24 July 2026

01 Image sources

Traceable origins

For every image used editorially, we check and document its origin, licence and permitted context of use.

Our own and editorial images

Photographs, graphics and illustrations from our editorial teams are documented with their creator, consent and, where applicable, model or property releases.

Agencies and stock media

We use stock images, including Adobe Stock material, within the licence purchased. Licence records and asset information are retained internally; we credit sources where required by the licence or editorial context.

Historical and open collections

Public-domain works, archives and open collections are credited with the institution, reference number or permalink where available.

Scientific illustrations

Tables, charts and figures from specialist contributions are published with source references and only within the rights granted or the limits of legally permitted quotation.

02 Artificial intelligence

Recognisable, not deceptive

For us, AI is a tool. Editorial responsibility always remains with people.

Our labels

“AI-generated illustration” identifies images that are wholly or predominantly generated.

“AI-edited illustrative image” identifies images whose main subject or meaning has been substantially altered by AI.

We do not separately label production adjustments such as cropping, noise reduction, colour correction or removal of small distracting elements where the subject, meaning and documentary character remain unchanged.

We do not use AI images as supposed evidence of real events. For healthcare topics, we also check for anatomical errors, misleading treatment situations, discriminatory depictions and potential confusion with real people.

Content of public interest and realistic manipulations are subject to the transparency requirements of the EU AI Act. Editorial exceptions apply only where human review and editorial responsibility are genuinely ensured.

03 How we work

Five review steps

  1. Check rights: Licence, copyright, personality rights and context of use.
  2. Document the source: Provider, asset ID, licence record, creation tool and editing.
  3. Review the content: Factual plausibility, medical safety, discrimination risks and potential deception.
  4. Ensure transparency: Source credit, illustrative-image label or AI label directly on the image or in the article.
  5. Take responsibility: Editorial approval, correction of errors and traceable handling of reports.

04 Existing material

Existing images

Older images with valid licences can generally continue to be used. New transparency rules do not retrospectively invalidate an existing licence.

We review older digital publications on a risk basis, prioritising realistic images of people, medical situations, content of public interest and images of unclear origin. We add missing or imprecise labels where the origin or editing method requires them.

For Adobe Stock material, we also consult the asset and licence information recorded at purchase. An asset being AI-generated does not automatically change its purchased usage licence, although editorial transparency may still require a label.

Reports and corrections

Is a source or label unclear?

Please identify the page, contribution and image. Our editorial team reviews reports and corrects identifiable errors.

Contact the editorial team

Further reading

This page describes our editorial practice. It does not replace legal advice and is updated when legislation or our workflows change.