Does Joshlilj use AI-assisted imagery?
AI-assisted image editing is identified in parts of this website’s photo information. For example, the biography photo information describes one portrait background as AI-assisted. That background should not be treated as evidence of a real location.
For current and future work, a material AI or synthetic alteration should be disclosed where the audience encounters the content when the alteration could affect how it is understood. An AI-generated or composite scene must not be presented as a documentary photograph of an event that did not occur.
An image-specific disclosure does not establish how every other photograph, video or social post was produced. The important questions are what was created or changed, whether that was clear to viewers, and what evidence supports the accompanying story.
How is a recorded event different from an illustration?
Footage of an interaction, an edited portrait and an illustration serve different purposes. A created scene cannot establish that the depicted event occurred, even when it looks like a photograph.
Recorded photograph or video
Camera-origin material. Its caption, editing, date and surrounding context still need to be accurate.
Ordinary editing
Adjustments such as cropping or exposure. An ordinary tool can still create a misleading impression through selection or omission.
Restored image
An existing image repaired or enhanced. Reconstructed details should not become new historical evidence.
AI-assisted or composite image
Generated, replaced or combined elements. An AI-assisted background, restoration or other material edit should be described accurately rather than presenting the result as untouched documentation.
Fictional or illustrative material
Not documentation that the people, quotations or events occurred as described.
A label describing an image does not verify its accompanying narrative. A fictional story with an AI label is still a fictional story. It must not be presented as an actual recipient case or evidence of delivered assistance.
What should a disclosure explain?
A disclosure should identify the material alteration where people encounter the content when that alteration could affect how the content is understood. It should not be available only in an image’s metadata, alternative text or a separate policy page.
For example, an altered background can be described as an AI-assisted background, while an invented scene needs to be identified as an illustration or fictional depiction. A generated quotation, voice or statement must not be represented as words a real participant actually spoke.
On social platforms, relevant native AI or altered-content disclosure tools should be used where applicable. A website policy does not replace a platform’s disclosure requirements.
What makes a factual statement supported?
A video, a participant’s account, a campaign total and an organization’s confirmation can support different facts. The wording should identify what kind of evidence is available rather than make every statement sound independently verified.
A participant can describe an experience without every detail having external corroboration. A campaign can display a total without establishing the amount received or a completed outcome. A published article supports what its reporting actually establishes, not every claim about the wider project.
- “The participant said...” for an attributed account.
- “The campaign displayed ... on [documented date]” for a dated platform amount.
- “The organization confirmed...” only when that confirmation exists.
- “A later outcome is not publicly established.” when the record does not establish one.
A machine-generated summary, transcript or search answer is not independent corroboration. It should be checked against its underlying source before being used as factual evidence.
Read how stories and sources are reviewed. See how media sources are classified. See how financial categories are distinguished.
What changes when someone appears again?
Previous participation does not automatically establish permission for a new recording, a materially different use or a new disclosure of personal information. Current wishes and the participant’s ability to make a meaningful choice remain relevant for new content involving recurring participants.
This current standard does not establish what permission was given in any individual historical encounter.
What if a caption, image or story is inaccurate?
A concern can be raised through the existing corrections route. Identify the content, the disputed detail and the requested change, without sharing unnecessary sensitive records.
A correction changes inaccurate information. A later update adds newly documented information. An illustration label identifies created material. These actions do different jobs: adding a label does not by itself correct an unsupported factual story.
Family members may raise concerns about older or reposted footage, including after a participant’s death. Those requests are reviewed individually. This does not create an automatic family veto or guarantee removal.
Where should someone check a particular story?
Start with the specific story and the records it cites. Compare the source date, what was actually documented and any later update. A general editorial policy cannot establish what happened in an unrelated interaction.
What these standards do not establish
These current and future standards do not resolve the origin, labeling, correction history or factual status of older social content. They also do not decide an incident-specific permission, family, financial or participant concern without the records for that matter.