Generative Artificial Intelligence (AI) could revolutionize healthcare, for example by facilitating the development of medicines or accelerating the detection of diseases, but the WHO believes that more attention needs to be paid to the risks.
In a document published this Thursday (18), the World Health Organization analyzed the dangers and benefits of using large multimodal models (LMM), a type of generative AI technology on the rise in healthcare. .
These LMMs can utilize multiple types of data, including text, images, and videos, and generate results that are not limited to the type of data entered into the algorithm.
The organization defines five areas that could use the technology: triage, to respond, for example, to written requests from patients; scientific research and drug development; for medical and nursing education purposes; administrative tasks; and use by patients, for example, to analyze symptoms.
Behind the great potential, however, the WHO warns that LMMs can provide false, inaccurate, biased or incomplete results.
“As LMMs are increasingly used in healthcare and medicine, errors, misuse and therefore harm to people are inevitable,” the WHO said.
– Technological giants –
The document also presented new guidelines on ethics and management of LMM, with more than 40 recommendations for governments, technology companies and health plans on how to benefit from this technology safely.
“Generative AI technologies have the potential to improve health care, but only if those who develop, regulate and use these technologies identify and fully take into account the associated risks,” says WHO Chief Scientist Jeremy Farrar.
The organization calls for the creation of liability standards to “ensure that users harmed by an LMM are appropriately compensated or have other forms of recourse.”
The organization also questioned whether LMMs were complying with existing regulations, especially regarding data protection.
Furthermore, the fact that large multimodal models are often developed and implemented by technological giants is worrying, according to the WHO, which recommends the participation of healthcare professionals and patients in the process.
Another point mentioned by the WHO was the vulnerability of these systems to cybersecurity risks, which could compromise patient information and even the security of medical care.
Finally, the organization concludes that governments should task regulatory authorities with approving the use of LMMs in healthcare, and calls for audits to assess the impact of this technology.
The post AI and health, a successful combination… or not, warns the WHO appeared first in ISTOÉ DINHEIRO.
#health #successful #combination #warns
