Going against the hype: looking at AI critically

EBHC Conference 2026 is approaching very fast. It will be, as conferences usually are, a time to see colleagues, friends, and meet new people. And that’s the keyword of the conference: people. Even when the event is heavily focused on artificial intelligence (AI), it is the people who matter.

We have been in this digital transformation for some time now, with paper being less and less a part of our health researchers’ lives. We do everything digitally: our lab diaries, our datasets, our documentation—all digital. We meet online, we record everything digitally, we use analytical tools that are only available because they are digital. Computers are really a wonder, and have revolutionised the way we research in health and public health.

The digital also offers many opportunities for us to practice open science: as open as possible, as closed as necessary, we say. If you have met me, you know how much I advocate for open science, complete reporting, transparency in research, and sharing everything —from methods, techniques, and data to equipment, tools, and research outputs. I am for the open and the open is possible because digital makes everything accessible for most.

But AI is a bit more recent in our lives. And there has been a lot of hype. I mean, a lot. The last four or five conferences I have been to have had 70% of their agenda taken by topics like “how to use it”, “why use it”, and “you will be out of the market soon if you don’t know how to use AI”. It is daunting, as the fashion now is to tame ChatGPT, to save time asking CoPilot to fix this code or correct this spreadsheet. If you haven’t been using it, you may be feeling excluded, not part of the conversation. FOMO – fear of missing out.

Well, at EBHC Conference you will have the opportunity to catch up — but I will be a bit annoying here and ask you to be cautious, to be critical, to be brave to question, to ask, to challenge. Whatever the conversation you have or the presentation you attend during the conference, I invite you to reflect:

• About time - How much time are you really going to save by asking AI to do a task for you? Will it get it right the first time, or will you have to prompt, reprompt, and adjust the prompt three, four, or ten times? Who is learning with that?
• About the environment - things we think are just simple, and which we have been doing for ourselves for our whole lives, now we ask AI to do for us. We know how to do it, but we only want to save time and effort. But each prompt to AI requires a huge computing effort, which in turn requires cooling of massive server facilities, consuming water and energy. Is it worth it?
• About your own intelligence - How much are you honing your own research skills when you ask ChatGPT to fix an Excel formula instead of searching and doing it yourself? How much are you learning, and how much are you delivering learning (for free) to the AI company?
• About safety - We forget this often, but each and every bit of information we feed AI tools with can tell them about us. Some AI tools now know a lot about me because I have been sharing information about my life with them. What happens if, instead of feeding our lives, we feed them our patients’ stories and data? How safe is it to upload a dataset to AI, and who is making a profit with that information?

These are only four concerns, among many others, that have been bothering me for some time. There are many more — and this comes from someone who has been working with AI and about AI for some time now.

So this will be a conference on evidence-based health care, held for over 20 years and proudly standing on the shoulders of leaders like David Sackett and others. These were brilliant people who made this conference happen for so long. So let us not forget that this is a collective construction of knowledge and skills, which requires, well… human intelligence. Let us not forget that. The people.