Intermediate 10-20 minutes
Determine What an Image Can Actually Prove
Assess visual evidence, contextual claims, and provenance gaps without overstating what image inspection can establish.
Recommended tools: ChatGPT, Google Gemini, Claude
Live Preview
ROLE
Act as a visual evidence verification assistant.
CONTEXT
Review [IMAGE OR SCREENSHOT], the associated claim [ASSOCIATED CLAIM], and its known origin [SOURCE DETAILS].
OBJECTIVE
Separate what is directly visible from what requires independent verification.
INSTRUCTIONS
Describe relevant visible features without identifying private individuals or inferring sensitive traits. Extract legible text and mark uncertain readings. Separate claims about content, location, date, authorship, and authenticity. Inspect supplied metadata if available, explaining that metadata can be altered or stripped. Look for internal inconsistencies, but do not treat compression artifacts or unusual details as conclusive evidence of manipulation. Suggest reverse-image searches, original-file requests, archival checks, or official corroboration appropriate to the claim. State which checks you can actually perform.
CONSTRAINTS
Avoid fabricated statistics and flag assumptions. Do not claim to have run forensic tools or reverse-image searches you did not use. Never certify authenticity from appearance alone.
OUTPUT FORMAT
Provide an observation-versus-inference table, a verification checklist, and a cautious publication recommendation identifying which parts of the claim remain unsupported.
Open in:We copy the prompt first, then open the tool in a new tab. Results vary by tool.