MJRC’s 2024 study remains the empirical foundation of this work. The new page does not replace it: it converts the project into a current company registry that can be updated as companies grow, fail, merge, change ownership or enter journalism workflows.

The categories follow what the systems actually do in the information chain. A company can span several functions, but GMFM assigns one principal category to keep the registry searchable and avoid double-counting.

Search by company, product or function. Existing public GMFM work links directly to the relevant profile; the remaining companies are mapped into the research universe and route to the central data-request system.

Some of the most influential AI systems are operated by companies already mapped in GMFM’s Tech Companies section. They matter enormously to journalism, but counting the parent company in both registries would distort the map.

Private AI companies are often much harder to examine than listed media groups. The page therefore makes the research layers visible without pretending that every underlying filing, investor record or structured dataset should be published openly.

GMFM does not treat company marketing material as financial evidence. Company documentation is useful for products and legal roles; ownership and financial claims are cross-checked against registries, securities filings, transaction documents, regulators and other primary records wherever possible.

This page is narrower than a generic AI-industry directory but broader than a list of chatbots. Inclusion depends on a material connection to journalism, public information or the production and verification systems news organisations increasingly use.