How Johannesburg's Visual Archives Ended Up Full of Ghosts: The Duplicate Image Problem Explained

The City of Johannesburg's digital asset libraries contain hundreds of thousands of photographs. A growing number of them appear more than once — sometimes dozens of times — under different file names, different dates and occasionally different rights clearances. That is the short version of how the municipality, the Gauteng provincial government and several affiliated cultural institutions arrived at what archivists now describe as a structural crisis in image management, one that predates the current ANC-DA coalition government but has landed squarely in its lap.

The problem matters now because the city is spending money. Several active procurement rounds — including a R42-million visual communications contract tied to the Joburg Development Agency's 2025/26 budget cycle — require clean, rights-verified image libraries. Duplicate images are not merely a storage inconvenience. Each one represents a potential copyright liability, a mis-spent licensing fee or, in the worst case, a photograph of a private individual that has been redistributed without consent under an incorrect Creative Commons tag.

Where It Started: The Digitisation Rush of the 2000s

The roots of the duplication problem trace back to roughly 2003, when both the City of Johannesburg Metropolitan Municipality and the South African Heritage Resources Agency began large-scale drives to scan physical photo archives. The goal was preservation. The execution was inconsistent. Different departments — the Johannesburg Roads Agency, the City Parks and Zoo administration on Asmara Avenue in Saxonwold, and the Newtown Cultural Precinct management — each ran their own digitisation projects with minimal coordination. Files were ingested into separate servers, renamed according to different internal conventions and, critically, never cross-referenced.

By 2010, the Soweto-based Avalon Creative Archive, which holds thousands of images documenting life in Meadowlands, Diepkloof and Orlando East from the 1970s onward, had donated a batch of roughly 8 000 scans to the City's central repository on Loveday Street in the CBD. At least 2 300 of those images were already in the system via a separate Heritage Month project. Neither the city's IT department nor the receiving archivist flagged the overlap at the time, according to a 2022 audit summary published by the Gauteng Department of Infrastructure Development.

The Johannesburg Art Gallery in Joubert Park ran into a version of the same problem when it digitised its photography collection between 2014 and 2017. Metadata standards used by the gallery's contracted digitisation firm did not match those used by the South African Broadcasting Corporation's image feed, which the municipality had also licensed. Result: the same images existed in the city's systems as both freely licensed municipal assets and as rights-reserved SABC property — simultaneously.

What the Coalition Found When It Took Over

When the ANC-DA coalition government in Gauteng began a broader audit of provincial digital infrastructure in early 2025, image library integrity appeared on the list largely because of a procurement dispute at the Gauteng Film Commission, headquartered in Parktown. An internal review found that at least one supplier had been paid twice for the same batch of stock photographs — a billing error that would have been caught automatically if a deduplication system had been in place.

The city has since contracted the Council for Scientific and Industrial Research, based in Pretoria, to develop a standardised hashing protocol that would flag identical or near-identical image files across municipal databases before they are ingested. The CSIR engagement began in March 2026, with a target completion date of December 2026.

For organisations working with Johannesburg's image libraries — community journalism outlets, urban planning firms submitting proposals to the Joburg Metropolitan Planning Tribunal, tourism developers pitching projects in the Maboneng Precinct — the practical advice is the same regardless of where the institutional failure originated: do not assume a photograph sourced from a city portal is rights-cleared. Cross-check the file name against the South African Institute of Intellectual Property Law's guidance on orphan works, verify the metadata independently, and keep a paper trail of every licensing decision. The deduplication work is ongoing. The liability, for now, is not.