Summary

This Computer Law & Security Review article examines investigative journalists' use of facial recognition technology in OSINT work and argues that the workflow raises a different risk profile than ordinary newsroom automation. The paper accepts that face matching can materially help identify people in images and video, but says publication, privacy, data-protection, and presumption-of-innocence risks scale up with the speed and reach of the tool.

Why It Matters

  • It is a direct journalism record because it focuses on how reporters and news outlets operationalize AI-assisted identity matching inside real investigations.
  • It is especially useful because it treats the workflow as a concrete reporting method, not just a generic ethics debate about AI.
  • It helps separate the investigative upside from the publication-stage risk: finding a possible identity is not the same as responsibly naming or depicting a person.
  • It adds a strong legal-rights lane to the archive's journalism verification coverage.

Investigator Workflow

The specific investigator task is using facial recognition on open-source imagery or video to identify, narrow, or corroborate a person of interest during OSINT-heavy casework. The maturity is `ad hoc tool`: this is a repeatable step inside a larger investigation, but not a stand-alone platform. The source states the journalistic face-matching workflow directly; the private-investigator transfer is inferred, but it is concrete and not speculative.

What the Source Says

The article says facial recognition has become an emerging newsroom practice, especially in investigative and OSINT settings where reporters are trying to identify people from images gathered in conflict or accountability reporting. It draws on legal analysis, literature review, and interviews with journalists and practitioners, and uses the BBC Verify investigation into the October 2023 Supernova festival attack as a key framing example. The paper argues that automated face matching changes the scale and speed of identification, which can also widen misidentification, surveillance, and function-creep risks if newsrooms do not impose stronger checks before publishing a match.