The World Health Organization warns of the dangers of generative artificial intelligence in the health sector – News Augsburg, Allgäu and Ulm

Care robot “Grace” at the ITU “AI for Good” Summit in July 2023Photo: AFP/Fabrice Coufrini

The World Health Organization has warned of the dangers of using so-called generative artificial intelligence in the health sector. A WHO analysis published in Geneva on Thursday concluded that although new technology could be useful in various areas of medicine, if it is not developed with sufficient data, it can lead to misleading and potentially dangerous results.

So-called generative artificial intelligence is used to generate texts, images or videos, for example in programs such as ChatGPT. Programs can be fed different types of data, transform them, manipulate them, and create new data from them.

The World Health Organization expects obstetric AI to be widely used in medicine in the future. The technology can be used in diagnosis, research and development, and in the education of doctors and nurses. The World Health Organization sees additional potential for generative AI in managing health data and supporting patients who want to know their symptoms.

According to the analysis, the danger of using generative artificial intelligence lies in the quality of the data on which the models are trained. Non-representative data also distort program outcomes, which may lead to incorrect, inaccurate, biased, or incomplete diagnoses.

The World Health Organization warned that the consequences in the health sector are equally serious. As the use of new technology in the health sector increases, “errors, misuse and ultimately harm to people's health are becoming unavoidable,” the analysis says.

The World Health Organization warns that the benefits of generative AI can only be harnessed if everyone involved uses the technology transparently and responsibly. To do this, rules must be set and medical staff and patients involved in developing AI programs for the healthcare sector.

“Generative AI has the opportunity to improve healthcare,” said Jeremy Farrar, chief scientific officer at the World Health Organization. “However, this can only be achieved if developers, decision-makers and users know the risks and feel responsible for them.”

The World Health Organization sees, among other things, a duty on big technology companies that develop software for the health sector. In addition, the risks of cyber attacks must be taken into account when handling sensitive health data.

The World Health Organization concluded its assessment of the opportunities and risks of generative AI in healthcare by calling on governments to regulate the use of the new technology and to regularly review it.

© 2024 Agence France-Presse