Ir al contenido
Do you give the green light for using AI in healthcare?

Do you give the green light for using AI in healthcare?

In a recent discussion, Nuria Pastor emphasized the importance of integrating artificial intelligence (AI) in the healthcare sector while adhering to certification standards and clinical evidence to ensure patient safety. She pointed out the risks of self-diagnosis using tools like ChatGPT, which, although valuable, lacks medical certification and is not designed for diagnosis. Pastor stressed the necessity for high-quality, reliable data to avoid creating biased AI, which could endanger lives. She asserted that medical startups must meet essential certification requirements without becoming overly burdened by them, highlighting the role of leadership, team management, company values, communication, and career planning in navigating this complex landscape. According to Nuria Pastor, proper governance is crucial for the successful application of AI in healthcare. Don’t miss more tips and insights on AI in healthcare in the full episode.

2025·Season 1·Episodio 4Highlight
Watch full episode

Nuria Pastor · CEO & Co-founder

Liderar con propósitoMejorar procesos

Watch the full "The Unboxing Talks" Serie

Transcription

00:05

Applying AI in the healthcare sector

00:12

Traffic light Green light for AI in the healthcare sector?

00:19

I think so.I believe that when things are done considering certification requirements and from a clinical point of view, with evidence to support it, we are not that far off.

00:29

But it is also true that not everything works that way.

00:31

In the end, if you make a clinical decision that is wrong,we are talking about a person's life, and we are talking about putting that person at risk.

00:37

That is why there are medical agencies,with whom we have obtained the CE marking and other types of agencies that have to ensure product quality.

00:50

What risks would you say there are to self-diagnosis using ChatGPT?

00:55

Well, ChatGPT is actually a very good tool.The thing is, of course, it's not certified as a medical device.

01:04

It's true that just two years ago, before it went viral,In Boston, people on the Microsoft board were already talking about it.

01:11

When will we see a ChatGPT with a medical device?But it's not designed to diagnose, and I think we still need to There's still a little time left for that.

01:21

But there are other AI solutions that are not far off.

01:25

What would be required in terms of patient-generated data so that AI could indeed provide all the valuable and accurate information such as that which a company like yours can offer?

01:37

I believe that data must be of high quality.Ultimately, the most important thing is that the data comes from reliable sources.

01:42

If you create AI based on an unreliable or biased source,or that does not take diversity into account, that does not take different genders into account,different variables that are a reality,you are creating contaminated AI and you could end up with a very dangerous AI.

01:58

That is why research and certification exists in the healthcare world,without it becoming an obsession.

02:06

Medical startups must be certified,but they don't have to be hyper-certified.

02:11

They have to comply with the certification they need, and that's it.