Arena’s Data Skills Co-ordinator, Adriana Homolová, outlines our data team’s stance on AI in the Dataharvest 2026 programme.
Large language models and generative AI have rapidly become an ever-larger presence in our private and professional lives. Some embrace the technology and use it daily; others are deeply critical and refuse to engage. Most of us are somewhere in between.
As journalists, being critical is second nature. And there is a plethora of important issues that come with this technology: privacy concerns, the extraction of value from our work, an insatiable energy hunger. It pollutes the internet with slop. It creates bad music.
Yet AI is also an amazingly powerful tool. Over the last ten years, I have seen many journalists struggle to achieve mastery in coding to make their stories better. For the first time ever, we don’t need years of training to write a piece of code. We still need to be able to instruct the AI clearly, but we no longer need to familiarize ourselves with for loops, conditionals or functions to get the work done. It can serve as a brainstorming partner, a first reader, a translator between an idea and its technical execution.
These two stances are not mutually exclusive. We can be critical of the macro-issues around AI technology, and still use it thoughtfully. What is important is that we create verifiable processes, systems of checks and balances, develop intuition and put down ethical guidelines for usage in the newsroom.
In a recent post on Columbia Journalism Review, How Journalists Can Make AI Work for Them, Stephen J. Adler concludes that while AI should be used carefully, disregarding it “would be like ignoring the internet at the turn of this century—or electricity in the previous one.”
At Dataharvest, we agree. We choose to educate instead of avoid. Nearly a quarter of this year’s sessions mention AI or large language models, and half of those are in the data skills track.
At Dataharvest 2026, we invite you to come and see for yourself what is hype and what is actually useful. You can learn about how large language models (LLM) can help you with FOIA requests, automate OSINT tasks or how to use it when working with sensitive information.
While most of our data skills sessions are for an intermediate audience, thanks to LLMs you can now also learn how to scrape or code basically anything, even as a complete beginner.
Take a look at the Dataharvest programme here.