Risks and Benefits of Medical AI
By Greg Kahn
Emerging Tech Exchange
Founder & CEO
Published on May 22, 2023
Machine learning has the potential to revolutionize healthcare by providing faster and more accurate diagnoses. However, there are concerns that AI could also be misused, leading to unintended consequences. The World Health Organization (WHO) is warning about the dangers of bias, misinformation, and privacy breaches when deploying large language models in healthcare.
WHO officials highlight the risk of misleading or inaccurate information generated by datasets that do not fully represent the population. Research conducted by the WHO suggests a 1 in 300 chance of harm occurring during a patient's journey, often due to data errors. It is crucial to recognize that AI in healthcare can both save lives and pose risks.
AI has already shown promising results in areas such as vaccine development and improving treatment for heart conditions. It is poised to aid in tackling challenging cancers and increasing the success rates of in vitro fertilization. However, disaster can be just one click or security breach away.
There are specific concerns regarding the accidental release of synthetic organisms and the potential for another global pandemic. Synthetic biology companies worldwide are working on creating artificial organisms, which could inadvertently give rise to antibiotic-resistant superbugs. The United Nations estimates that superbugs could cause 10 million deaths annually by 2050, surpassing cancer as a leading cause of death.
Furthermore, the misuse of AI models can have serious implications. Researchers have demonstrated the ability to create thousands of new chemical weapons compounds in a short time using AI models. Additionally, large language models in healthcare can generate false information, known as hallucinations, which could be harmful in medical settings.
Bias in AI models is another significant concern, as it can perpetuate racial, gender, and geographic disparities in healthcare. Access to technology and equal opportunities must be ensured to avoid exacerbating these disparities.
The current regulatory frameworks, such as the FDA's approach to medical device regulation, are not fully equipped to handle the influx of AI-powered healthcare applications and devices. Updated guidance from regulatory authorities is needed to address the specific challenges posed by AI in healthcare.
Ultimately, while AI holds immense potential in transforming healthcare, the risks are something we have to nail down now. Striking a balance between innovation and safeguarding patient well-being is of paramount importance. Regulatory measures, transparency in algorithms, and comprehensive guidelines are necessary to harness the benefits of AI while mitigating potential harms.
Greg Kahn
Emerging Tech Exchange
Founder & CEO
