WHO warns of inevitable misuse and harm as it highlights AI risks in healthcare.
The World Health Organization emphasized on Thursday that while generative artificial intelligence has the potential to revolutionize healthcare by enhancing drug development and expediting diagnoses, it is crucial to prioritize the associated risks.
WHO has been investigating the potential dangers and benefits of large multimodal models (LMM) in artificial intelligence. They are relatively new and quickly adopt to health.
LMMs are a type of generative artificial intelligence that can use multiple types of data input, including text, images, and videos, and produce outputs that are not limited to the type of data input to the algorithm.
“It is predicted that LMMs will have wide use and application in healthcare, scientific research, public health and drug development,” said the WHO.
The UN Health Agency outlined five broad areas where the technology could be applied.
These include: diagnosis, such as answering written patient questionnaires; scientific research and drug development; medical and nursing education; office duties; and patient-directed use, such as examining symptoms.
– Abuse, damage ‘unavoidable’ –
Although this has potential, the WHO cautioned that there are documented risks that LMMs may produce false, inaccurate, biased or incomplete results.
They may also be educated about low-quality information or information that contains biases related to race, ethnicity, ancestry, gender, gender identity, or age.
“As LMMs become more widely used in health care and medicine, errors, misuse and ultimately harm to individuals are inevitable,” the WHO warned.
On Thursday, it released recommendations on the ethics and governance of LMMs to help governments, technology companies and healthcare providers use the technology safely.
“Generative AI technologies have the potential to improve health care, but only if those who develop, regulate and use these technologies recognize and fully consider the risks involved,” said Jeremy Farrar, WHO’s lead researcher.
“We need transparent information and operating methods to manage the design, development and use of LMMs.”
The WHO said liability rules are needed “to ensure that users harmed by LMM receive adequate compensation or have other legal remedies”.
– The role of tech giants –
Artificial intelligence is already used in diagnosis and clinical treatment, for example in radiology and medical imaging.
However, the WHO emphasized that forms of LMM carry “risks that societies, health systems and end-users may not yet be ready to fully consider”.
This included concerns about whether LMMs were compliant with current regulations, including data protection – and the fact that they are often developed by tech giants as they require significant resources and could therefore strengthen the dominance of these companies.
The guidelines recommended that researchers and engineers should not develop LMM methods alone, but also medical professionals and patients.
The WHO also warned that LMMs are vulnerable to cybersecurity risks that could compromise patient information or even the reliability of healthcare.
It stated that governments should allow a regulatory authority to approve the use of LMM in healthcare, and there should be audits and impact assessments.
Also read these top stories of today:
Samsung vs Apple! Samsung introduced the Galaxy S24 series with AI features to create a dominant position in the market. Could it reclaim the smartphone throne after losing to Apple in 2023? See how Samsung plans to beat Apple here.
Samsung Galaxy S24 released! Samsung’s latest flagship smartphones were unveiled with updated features at the Galaxy Unpacked event in California. It goes hand in hand with Apple’s iPhone 15. Check out the technical comparison between Samsung Galaxy S24 and iPhone 15. Know how it differs from its top competitor here.
“Just the right” robots! Moxie is the size of a teddy bear and powered by generative artificial intelligence, and can interact with children by telling them stories, giving math lessons and performing dance moves with two hands. Want to know more? Read here.
