Using personalisation makes every AI chat less risky and more effective — here are five essential custom instructions to add

Leave a reply
5 personalisation settings for genAI tools Reduce sycophancy Tell AI to question assumptions. Set your context Add your role, audience, values and goals. Choose your style Set language and tone, and require succinctness. Reduce deskilling Ask AI to coach, explain and build your thinking. Allow uncertainty Set thresholds for quality and context.

One of best things you can do to lower the risks of using AI and improve the quality of outputs is to spend five minutes with one of AI tools’ least-used features: their personalisation settings.

ChatGPT, Gemini, Copilot and Claude all have personalisation features: a section of the tool where you can give it permanent extra information that informs every conversation. You can find it in your general settings under names like ‘instructions’, ‘personal context’ or ‘personalization’, or using the links above.

Here are five instructions you can include in that section:

Continue reading →

How to judge (and minimise) the risk of using sensitive information with an AI chatbot

Leave a reply

What are the risks of information you put into an AI chat becoming public? Evaluating those risks is slightly different to other information security challenges because of the way large language models work, so here’s a guide to assessing and managing information security with AI.

4 security questions for genAI Uniqueness How much does the LLM already know about this? Sensitivity How bad would disclosure be? Exposure Who or what could access it besides the model? Alternatives Are there safer ways to complete the task? Paul Bradshaw 2026 | Icons generated by Google Gemini
Continue reading →

How data journalism is done on TikTok: six key takeaways from watching 50 videos

Leave a reply

In a guest post for OJB, James Hickman pulls together a set of basic best-practice guidelines for creating data journalism videos on TikTok, from length and location to animation and style.

6 best practice guidelines for video data shorts: 1. Keep your data shorts under 90 seconds 2. Maintain a human connection 3. Locations reinforce authority 4. Show don't tell 5. Animate your graphs 6. Don't be afraid to innovate and adapt

Short-form vertical video is quickly becoming a key format for telling all types of news stories — and data journalism is no exception.

As a data journalist, I wanted to know how these outlets use this new and evolving digital video format to create “data shorts”: short-form social media videos in a vertical, reel-style format, that analyses or reports on specific datasets or newly released data.

I analysed 50 data shorts from the TikTok accounts of five major UK news publications — the BBC, ITV News, Guardian, Times and Financial Times — to examine how these digital news outlets approach the format. I looked at the genre of story they covered, the angles they focused on, their length, and their use of animation, to help inform my guidelines.

Continue reading →

This one story can be used to discuss seven different types of bias

Leave a reply
ITV News headline: John Torode’s wife Lisa Faulkner reveals breast cancer diagnosis

The latest “wife of” headline — ITV News’s report on the actor Lisa Faulkner revealing that she has undergone surgery after a cancer diagnosis — is an opportunity to get journalism students exploring how different forms of bias might shape news reporting — and not just the obvious ones.

Continue reading →

Managing a mass FOI project? Here’s an AI-assisted methodology for that

Leave a reply

Sending FOIs to multiple bodies across the country to get the big picture on an issue sounds like a great idea — until the responses start to trickle in. Differences between responses often make mass FOI projects extremely time-consuming as you try to get everything into a format that allows you to ask journalistic questions and compare different authorities. Can AI help?

On one recent project I decided to put together a methodology that made the process less stressful, faster and more accurate. Here’s how it works.

Data structure Extract & reshape Check & verify Combine Audit & prioritise Audit responses to identify the level of detail in each response and identify edge cases. Include a caveats column. Augment manual audit with NotebookLM audit. Identify a priority order for data, e.g. totals by outcome, hospital, category or year where these are provided separately Design a data structure that can accommodate all responses Structure should follow ‘tidy’ data principles, i.e. one row per combination of features (force, category, hospital, outcome, year) Structure should include source details, e.g. filename, sheet name, name of person entering data PDFs: use Tabula or vibe coding (design a prompt template to generate code to attempt to extract data). Multi-sheet XLS files: use Open Refine to import and combine sheets Design a prompt template for generating code to reshape CSV responses Manual checks (e.g. compare entries, check page-ending rows) Analysis-based checks (e.g pivots, totals) AI-based checks using a prompt template (e.g. compare files) Use OpenRefine or: Design a prompt template for generating code to combine the resulting CSV files.
Continue reading →

Showing charts on video? Here are two essential techniques to make them effective

Leave a reply

Using visualisation on TV and video is very different to using charts and maps online. In video, the audience has very little time to absorb the information contained in the chart — so you need to get them to that information as quickly as possible.

Every bad example of charts in videos forgets this. And every good example uses two essential techniques: keeping things simple, and adding motion.

Continue reading →

“Many newsrooms are not optimised for what humans do best” (but we have an opportunity to change that)

Leave a reply
Amazon worker with horse head
Image by Cory Doctorow, from Revenge of the Chickenized Reverse-Centaurs

Some essential reading by Agnes Stenbom Swedling explores how news organisations integrate AI into their workflows and the idea of the “human in the loop“. Many newsrooms, she points out, “are not optimised for what humans do best”, and so far the introduction of AI hasn’t involved a critical consideration of whether we want to embed those features in new systems, or rethink them:

“What is being built – incrementally, often unintentionally – is a form of machine-centric hybridisation. Workflows are optimised for what machines do well: speed, scale, pattern recognition, cost efficiency. Humans are then positioned around those systems, adapting their tasks, roles, and decision-making to fit the logics of machines. 

“The consequence is a subtle but significant inversion: rather than engaging in uniquely human activities, work is reorganised to fit machine-driven processes. And once that inversion is embedded at the infrastructural level, it becomes increasingly difficult to reverse.”

Continue reading →

Caught in a trap: what journalists can learn from systems thinking

Leave a reply

One of the most powerful ways to generate original journalism is to look at the systems behind stories — particularly the points where those systems fail.

For investigative work, those points are central. Surface-level scandals often stem from deeper systemic problems. So what tools do we have for recognising those patterns?

Donella Meadows’s classic book Thinking in Systems offers one: “system traps” — patterns that explain how systems get stuck, break down, or behave in ways nobody intends. They are “traps” because attempts to escape them often backfire.

System trap Journalism examples Policy resistance The war on drugs; reforms that fail; missed targets Overuse leading to shortages; climate change impacts; AI Tragedy of the commons Drift to low performance Normalisation of poor performance or low productivity Escalation Arms races; races to the bottom Success to the successful Increasing concentration of wealth or resources Shifting the burden to the intervenor Subsidies, price fixes and delaying the impact/cost of a policy Rule beating Tax avoidance, loopholes Seeking the wrong goal Schools focusing on targets over pupil welfare;

In this post I’ll explain each trap, what it looks like in the wild, and how to use it as a lens for story ideas.

Continue reading →

How to: generate hundreds of maps by combining QGIS with Python (code included!)

Leave a reply

At this year’s Dataharvest I delivered a workshop on using Python in QGIS to automate the process of exporting maps for multiple locations. Here’s how to do it (you can find a GitHub repository with materials and links here).

Making a map for a story is cool — but what if you could make a map for every reader? Or if you’re working on a project involving teams in different regions or countries, what if you could give each one of those teams a map centred on their own patch?

Normally you would have to manually move the map to centre it on a key city, and then export an image. Then do it again and again and again for every area.

Luckily, QGIS has the ability to run code. And this is a great excuse to start using it.

By organising the layers on the left you can put shapes such as flood defences over a base OpenStreetMap layer. You can also change the scale in the box underneath the map
Continue reading →

FAQ: AI, misinformation and journalism

Leave a reply

In this latest post in the FAQ series, I am sharing some responses to a radio interview about AI’s impact on journalism.

Q: Is the continuous growth of AI-generated content online a danger for 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 reading →