Checking metadata in reporting notes, recordings, and photos

Checking Metadata in Reporting Notes, Recordings, and Photos

Reporting materials should not be judged by their visible text alone.

Notes, recordings, photos, videos, PDFs, and screenshots can retain reporting dates and times, locations, devices, creators, revision histories, filenames, background sounds, and visible details captured in the material.

Even if a source's name is withheld in the article text, metadata and surrounding information in the materials can allow the source or location to be inferred.

For journalists, metadata checking is not a technical detail.

It is pre-publication work for protecting sources.

This article organizes the information to check before publishing or sharing reporting notes, recordings, and photos.

Information left in reporting notes

Reporting notes are not just text.

Digital notes, collaboratively edited documents, cloud notes, and smartphone notes may retain creators, editors, creation times, update history, and sharing history.

Notes used for publication should be separated from originals.

Originals may be necessary as reporting records. At the same time, external sharing and public materials should not retain unnecessary creator information or comments.

Recordings contain more than voices

In recording files, not only spoken content but also background sound and file information become clues.

Station announcements, in-store music, workplace machinery, ambulances, school bells, surrounding conversations. These can all be material for inferring place and time.

When publishing audio, changing the voice alone is not always enough.

Also check background sound, wording, conversation context, and the length of the recording.

In high-risk reporting, there may also be a decision to switch to a transcript or summary instead of publishing audio as-is.

Look at backgrounds in photos and videos

In photos and videos, check not only faces but also backgrounds.

Signs at the reporting location, what is outside windows, documents on desks, name tags, uniforms, reflections, wall notices, and device screens. These can point to sources or locations.

When processing photos, it is important not to overtrust blurring or cropping.

Text can sometimes be read through weak blur. Information from before cropping may remain in another layer or preview.

Create a publication copy and open it in another app to check it.

Check with tools such as ExifTool

ExifTool is sometimes used to check metadata in images, audio, and video.

ExifTool is a tool that can check and edit many kinds of metadata, including EXIF, GPS, XMP, and IPTC. It is used not only for photos but also for checking information in a wide range of files, including video and audio.

URL : https://exiftool.org/

The reason to introduce ExifTool is that it is practical for checking metadata in reporting materials.

However, the information visible through a tool is not everything. Background sounds, visible details in images, distinctive expressions in text, publication time, and cloud sharing history must be checked separately.

Create a publication copy

Separate reporting materials into originals, checking copies, and publication copies.

It is important not to publish originals directly.

However, for materials where evidentiary value matters, careless processing can cause other problems. For materials involving legal judgment or public-interest whistleblowing, proceed in consultation with the newsroom, lawyers, and specialists.

Review storage locations and sharing methods

Even if metadata is checked, weak storage locations or sharing methods can break source protection.

If materials are placed in a personal cloud account, real-name account, workplace shared folder, collaborative editing tool, or messaging app, owners, viewers, editors, and access times may remain.

Store reporting materials after deciding who may view them.

Make them accessible only to necessary people, and do not casually widen shared links.

Transcripts also retain clues

Turning a recording into a transcript does not necessarily make it safe.

Transcripts retain speaking style, verbal habits, dialect, technical terms, names of people nearby, and the sequence of time. If you use an AI transcription service, you are handing the audio data and transcript content to an external service.

When publishing a transcript, the content itself also needs anonymization.

Using an external transcription service means giving the audio or transcript content to another party you must trust. For high-risk reporting, unpublished materials, or audio from which a source could be identified, prioritize an operation that does not send the material to external services. If such a service is used, check its terms of use, retention period, sharing scope, and handling of confidential information.

Separate review roles

The person who created the reporting material can easily miss things.

Background name tags, place names in audio, PDF creators, and screenshot notifications are sometimes noticed by a third party.

Inside the newsroom, safety improves when text review, file review, and source protection review are separated.

For high-risk reporting, it is safer not to complete material review alone.

Summary

Reporting notes, recordings, photos, and videos retain information beyond the article text.

Creators, revision history, recording time, background sound, GPS, capture time, filenames, and visible details captured in the material become clues pointing to sources or locations.

Before publication, separate originals, checking copies, and publication copies, and check both metadata and content.

Tools such as ExifTool help with metadata checks, but they do not automatically protect against background sound, visible details, distinctive expressions in text, or post-publication correlation.

Checking reporting materials is an important process for protecting sources.

Related tools

Related articles