Methodology and transparency
How the films are made
From topic to publication: peer-reviewed sources, medical review by the author, transparent use of AI and a public corrections log for every film.
A science documentary has to be engaging, but before anything else it has to be true. This page explains how every History of Infections film is made, where the data come from, what artificial intelligence does and does not do, and how we correct ourselves when we get something wrong.
The process, step by step
Every film starts from a concrete question: where a disease came from, why an epidemic was so deadly, which common infections can still kill. The question decides what needs researching and what stays out.
We look for peer-reviewed, recent, published and verifiable work. We use worldwide data and look closely at the places where each disease actually hits hardest, not only at the countries that make the headlines. In the rankings, every position rests on that data, not on an impression.
The sources become a script written for any curious viewer: no unnecessary jargon, real stories and real people, and figures rounded only when rounding does not change their meaning.
The author, an infectious disease physician, reviews the entire script: figures, mechanisms, treatments and nuances. No medical claim reaches the screen without that review.
Narration, archival images, illustrations, maps, music and editing. Further down we explain which parts of this step use artificial intelligence.
The film premieres on YouTube and its page goes up here, with a summary, the key points, the complete sources and, if there are any, the corrections.
What sources we use
- Peer-reviewed. The foundation is articles in peer-reviewed scientific journals: original studies, systematic reviews and meta-analyses.
- Current. For present-day figures (cases, deaths, antibiotic resistance) we prefer the most recent publication available. When a figure is several years old, we say so.
- Verifiable. Every source on a film's page carries its DOI or its link, so anyone can go to the original.
- Global. We give the worldwide picture and actively seek out data from Africa, Asia and Latin America, which are often left out of epidemic stories.
- Complementary, when needed. For historical events we also use books on the history of medicine and documents from the period; for surveillance figures, reports from the WHO and national public health agencies such as the CDC. They are cited just like the articles.
Why references do not appear on screen
In the films you will not see slides with article titles, authors, journals or DOIs, and the narration does not read them out. That is an editorial choice: a documentary needs pace, and stopping the story to read a citation turns it into a conference talk.
That does not mean the sources are hidden. They are listed in full on each film's page on this website, in the documentaries section, where you can look at them at leisure and follow the link to the original.
Artificial intelligence: what it does and what it does not
We use artificial intelligence tools and would rather say so plainly:
- The voice. The narration is the author's own voice, cloned with ElevenLabs from recordings of his real voice. It is not a stock voice or anyone else's.
- The images. Illustrations and some animated clips are generated with AI and are labelled as such in the films. Archival photographs and footage are real and are credited on screen.
- Research and production. AI helps to find and organize sources, prepare drafts, translate and edit the film. The medical content is reviewed and signed off by the author, who answers for it.
- On YouTube. The films are disclosed on the platform as altered or synthetic content, so YouTube can show that to viewers.
What AI does not do: decide what is true. Every figure is checked against its source and every medical claim goes through the author's review.
Corrections
Even with all that care, we can get things wrong. If you find an error, we would be grateful if you told us:
- Write to ivantrompa@historyofinfections.com with the subject line “Correction”.
- Tell us the film, the approximate timestamp, what it says and what it should say.
- If you can, include the source that supports the correction.
We review every report. If the error is confirmed, the correction is published on the film's page here, under “Corrections”, with the date and a short explanation of what changed. Corrections are never deleted: they stay in plain sight.