A former broadcast journalist built an AI tool that flags anti-Black bias
A fine-tuned GPT that spots anti-Black narratives in articles, built with Make.com after coding it the hard way stalled.
This former broadcast journalist, who worked at ITV News and BBC radio, founded a social enterprise to challenge how UK media portrays Black and other minoritised communities. In the program they found the missing piece to turn that mission into a working AI tool. It is a clear example of a founder using new skills to serve a cause they have carried for years.
Media Bias Checker Tool
Before
They had worked in newsrooms at ITV News, BBC radio and Unity 101 community radio, and went on to found Our Version Media to challenge the demonisation of Black and other minoritised communities. They wanted a tool that could catch biased narratives automatically. But they had been trying to build it the hard way, with code.
“I’d previously been trying to create the Tool the hard way - using code!”
In their own words
The turning point
During the Mastery course, Vykintas's two Make.com lessons showed them a different route. They learned how to connect integrations and APIs into a working system. That made it possible to fine-tune an OpenAI GPT and wrap it in a tool people could actually use.
“The Tool is powered by Make.com integrations and APIs. I learned the process during Vykintas’s two Make.com lessons.”
In their own words
After
They built the Media Bias Checker Tool, trained on a dataset of media-driven stereotypes, public reports from Our Version Media's Media Reporting Tool, data from its Media and Wellbeing Survey, and synthetic data. They added a user interface for a seamless experience. Early feedback has been positive, and they are now working on Version 2.0.
What it lets people do
Readers can check articles for anti-Black narratives, which makes reporting to press regulators easier. Journalists can check their own writing before they hit publish and learn to unlearn their biases. That makes it a learning tool for fair representation and racial equity too.
What they learned
The right method mattered more than the hardest one: a no-code approach got them further than code alone. Having just finished a Masters in Applied AI and Data Science, they found the Mastery course more relevant to their work as a founder, changemaker and social entrepreneur.
“Even as it is, this first version, it could help to change and reduce the harmful negative narratives that negatively impact people's lives, wellbeing and life chances.”
In their own words
Media Bias Checker Tool was built during AI Mastery.
Read the full submission
Written 20 January 2025 Submitted by the storyteller
Every story is written by a Mindvalley student or by someone our teachers have worked with. How we collect and check stories
More from AI Mastery
