In this latest post in the FAQ series, I am sharing some responses to a radio interview about AI’s impact on journalism.
It is certainly a problem yes, in three ways: it makes reporting harder, it makes it harder to support journalism financially, and it makes it harder for audiences to trust your reporting.
Continue readingFrom stories about political speech and song lyrics, to street names and social media chatter, data journalists now have a wide range of examples of text-as-data to draw inspiration and guidance from, while tools such as Pinpoint and NotebookLM are making text analysis easier than ever.
I compiled a list of over 200 pieces of data journalism where text or documents were used as sources. Quantification techniques ranged from counting the frequency of a single word and using Google’s ngram viewer, to machine learning and topic modelling.
Looking at those articles it’s clear that, once quantified, journalists tell the same stories about text as any other piece of data: using the seven most common angles.
But how those angles are used — and how often — is where it gets interesting…
PEER is a mnemonic (based on a previous post) for remembering the following four types of source:
Each type of source brings something different to the story: voices of power primarily (but not solely) answer questions about action: what was or is being done, what should or would be done about a particular issue. These are easily the most commonly quoted sources in news reporting.
People with expertise can answer the “why” and “how” questions — and are often more likely to speak to journalists — while those with experience can verify or validate (put a human face to) events. Representatives can speak to the wider impact or significance of an issue, or represent community sentiment about it.
Making each type of source explicit allows us to think about what those roles really mean — and identify less obvious ideas for sources with power, expertise, experience or representative qualities.
Continue readingFreedom of Information (FOI) requests are not only one of the best ways to get original and exclusive stories that set your reporting apart — they’re also a good way to develop core journalism habits like curiosity, scepticism, and creativity. Here are some tips on how to get started with FOI while developing those qualities.
Curiosity is the first quality I identified in my series on the 7 habits of successful journalists — and FOI is a great way to hone that.
One good way to get started with FOI is to identify an event or problem that you’ve read about, and get curious about it: how many times is that event happening? How much is that problem costing? These are perfect questions for FOI.
Continue reading
A pirâmide invertida do jornalismo de dados mapeia o processo de utilização de dados na reportagem, desde a geração de ideias, passando pela limpeza, contextualização e combinação, até à comunicação. A fase final — a comunicação — apresenta uma série de opções: desde a visualização e sonificação até à personalização e ferramentas. Mas quais são as melhores práticas para cada uma?
(Também disponível em inglês, alemão e espanhol, russo e ucraniano).
A visualização é normalmente a forma mais rápida de comunicar os resultados do jornalismo de dados: ferramentas gratuitas como Datawrapper e Flourish muitas vezes exigem apenas que você carregue os seus dados e escolha entre várias opções de visualização.
Continue reading
Os projetos de jornalismo de dados envolvem várias etapas, cada uma apresentando seus próprios desafios. Para ajudar a compreendê-las, criei o que chamei de ‘Pirâmide Invertida do Jornalismo de Dados’. Ela delineia as etapas que precisam ser consideradas à medida que a matéria avança desde a conceção inicial até a comunicação dos resultados, e como elas se relacionam entre si. Abaixo, explico cada etapa, identifico questões a considerar conforme o projeto avança e ofereço conselhos e dicas sobre como enfrentá-las.
(Também disponível em Inglês, Alemão, Espanhol, Finlandês, Russo e Ucraniano.)
O primeiro desafio que um jornalista enfrenta é conceber uma ideia viável para uma matéria baseada em dados.
Continue reading
“Parallel prompting“* is the term I use for this: while you perform a task manually, you also get the AI to perform the same task algorithmically.
For example, you might brainstorm ideas for a story while asking ChatGPT to do the same. Or you might look for potential leads in a company report — and upload it to NotebookLM to perform the same task. You might draft an FOI request but get Claude to draft one too, or get Copilot to rewrite the intro to a story while you attempt the same thing.
Then you compare the results.
Continue readingMarcus Bösch, the editor of the Understanding TikTok newsletter, has put together a checklist for “when a meme is everywhere and you’re unsure whether to cover it, contextualise it, or leave it alone.” (PDF version here).
The checklist — M.A.T.T.E.R. — covers six things to consider: Meaning, ‘Affect’ (emotion), Type of format, Temporality, Ethics and relevance.
In September I took part in a panel at the African Journalism Education Network conference. The most interesting moment came when members of the audience were asked if they didn’t use AI — and why.
Last month I wrote about destination and journey prompts, and the strategy of designing AI prompts to avoid deskilling. In some situations a third, hybrid approach can also be useful. In this post I explain how such hybrid destination-journey prompting works in practice, and where it might be most appropriate.
Continue reading
Online Journalism Blog 