14-01-2020 7:11 am Published by Nederland.ai 1 Comment

Artificial Intelligence (AI) is not new. For decades, people have been thinking about robots and computers that can think like humans. It will not have escaped anyone's attention that the AI developments over the past few years have gone incredibly fast. Self-driving cars and face recognition are examples of this, but developments in health care are also ongoing. This sector is ideally suited for AI because of the enormous amounts of data. And the growth is not over yet. But what is AI and what issues does it raise? The report ' Artificial intelligence in healthcare ' examines what AI is, applications in practice and what issues this raises.

Main categories of AI
Definitions of AI are not clearly described and cannot be explained in one sentence. Nick Bostrom, a Swedish philosopher who is affiliated with the University of Oxford, uses a classification in relation to  human intelligence. Almost all current AI applications fall into the Artificial Narrow Intelligence category (weak AI).

Typical for this category of AI is that it specializes in one specific area. Early
detection of dementia, predicting lung cancer, predicting the outcomes of a patient in Intensive Care, detection of abnormal cells in pathology and recognizing dermatological abnormalities are some examples of applications of

AI in healthcare covered by weak AI. These applications are not yet widely used. However, the potential is high, because AI at this level increases the quality of care and saves time for the caregiver and the patient.

Machine learning

Almost all current AI applications in healthcare make use of machine learning technology . Machine learning gives systems the opportunity to learn and improve themselves on the basis of previous experiences. Deep learning is a specific form of this technique. Deep learning is suitable for recognizing highly detailed patterns such as specific abnormalities on a CT scan.

Issues

AI is a promising development, but to be able to apply AI safely and responsibly on a large scale in healthcare, issues must be resolved. For example, data must be of sufficient quantitative and qualitative level for reliable results, even if they come from different sources. There are also ethical challenges: who is ultimately responsible for a decision when an algorithm is wrong?

More information

If you have any questions about the ' Artificial intelligence in healthcare ' report, please contact Jacqueline Nell (06 143 363 59, nell@nictiz.nl ). If you have questions about AI in general, you can contact Frauke Wouda, MD ( wouda@nictiz.nl ) and Henk Hutink, MBA ( hutink@nictiz.nl ).

Tags: , , ,

1 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

19 + 16 =