Sources and methods

The information on promptandmodel is compiled from a defined set of published news feeds and indexes from established publications. These sources cover the site's core subject matter: the development of AI models and the regulatory rule governing their deployment. Each news story from these feeds is processed with automated assistance to extract its key factual claims. It is then rewritten into a concise summary, always credited to the original publication from which it was sourced.

When a factual claim appears in only a single publication, it is treated as an unverified report. Such claims are not presented as established fact on promptandmodel. They may be included in summaries with clear attribution to that sole source, allowing readers to note the provisional nature of the information. The site's focus is on aggregating and cross-referencing reports to build a stable picture of developments.

The reference pages for models, regulation, techniques, companies, tools, authorities and benchmarks are built differently. They are constructed from facts that appear consistently across multiple reports or from official, stable sources like legislation or technical papers. Each reference page centres on a factual table and is accompanied by a credited photograph where one is available under licence. This creates an at-a-glance resource built from verified information.

All content on promptandmodel is subject to regular refresh cycles. News coverage is updated continuously as new reports are processed from the source feeds. The reference pages are reviewed and updated on a scheduled basis to incorporate new, stable facts and reflect significant, verified developments in the field. This ensures the site remains a current and reliable reference.