São Paulo's Digital Archives Contain Thousands of Ghost Images, Experts Warn

São Paulo's municipal digital systems are carrying thousands of redundant, corrupted, or duplicated image files, a quiet crisis in public-sector data management that specialists and city administrators say has grown too large to ignore. The problem spans everything from the Prefeitura de São Paulo's official procurement transparency portal to the digitised collections held by institutions along Avenida Paulista, and fixing it, experts warn, is neither cheap nor simple.

The issue has moved up the agenda partly because of a broader push by the Nunes administration to modernise the city's e-government infrastructure ahead of the 2026 municipal budget cycle. Digital governance teams inside the Secretaria Municipal de Inovação e Tecnologia have been auditing legacy content management systems since late 2025, and what they found, according to public procurement documents reviewed by this newspaper, includes image repositories where the same photograph or scanned document appears dozens of times under different file names, slowing load times and making keyword searches unreliable.

Why Duplicate Images Are More Than a Housekeeping Problem

At street level the consequences are mundane but real. Citizens trying to access building permits through the SP156 platform, or journalists pulling archived planning documents from the GeoSampa mapping service, frequently encounter broken image placeholders or are served the wrong version of a scanned file. GeoSampa, which covers the entire 1,521-square-kilometre municipality, relies on a geodata image layer that is updated by dozens of different municipal departments with no single deduplication standard in place.

Technologists working with the Associação Brasileira de Entidades de Tecnologia da Informação e Comunicação, known as Brasscom, have flagged the problem as symptomatic of how Brazilian public institutions adopted digital workflows rapidly without standardising metadata rules. The result is what one industry report from April 2026 described, without naming specific municipalities, as "content entropy": a state in which digital assets multiply faster than they are curated or removed. São Paulo, as the country's largest municipal government by budget and by the sheer volume of daily digital transactions, is a prime example.

The financial dimension matters too. Cloud storage costs for the Prefeitura have risen. Municipal budget documents published in the Diário Oficial da Cidade de São Paulo in March 2026 show the city allocated R$47 million to digital infrastructure contracts for the current fiscal year, a figure that includes storage, licensing and maintenance. Specialists in public-sector IT who spoke generally about the issue, without commenting on São Paulo specifically, estimate that unmanaged duplicate assets can inflate storage costs by between 15 and 30 percent in large government repositories.

Institutions and What They Are Doing About It

The Museu de Arte de São Paulo, the MASP, on Avenida Paulista, completed its own internal image deduplication project in early 2026 as part of a broader collections digitisation drive. The museum's publicly available project notes describe deploying hash-based image comparison software to clear redundant files from its digital catalogue, a technical approach that cultural heritage professionals elsewhere in the city, including teams at the Memorial da América Latina in Barra Funda, are now studying as a potential model.

At the federal level, the Lula government's Rede Nacional de Ensino e Pesquisa, the RNP, has been piloting AI-assisted deduplication tools across university digital libraries since 2025. Whether those tools can be adapted for municipal government use at the scale São Paulo requires is a question the Secretaria Municipal de Inovação e Tecnologia has not yet answered publicly.

For residents and businesses that rely on public digital services, the practical advice from digital governance specialists is consistent: when downloading official documents from city portals, always check the file's published date and document number against the original in the Diário Oficial, since a duplicate image file can carry an older version of an amended document. The city's SP156 helpline, reachable at 156, can flag specific broken-record complaints directly to the relevant secretariat. A formal deduplication policy, officials have indicated, is expected to be part of the next digital governance decree, anticipated before the end of the third quarter of 2026.