AI in Healthcare May Exacerbate Racial and Income Disparities

2023-07-24

Scientists are concerned that the use of artificial intelligence models like ChatGPT in the healthcare field will exacerbate inequalities. Epidemiologists from the University of Cambridge and the University of Leicester have warned that large language models (LLMs) could reinforce unfairness among racial minority groups and low-income countries. Their concerns stem from systemic data biases. The AI models used in healthcare are trained based on information from websites and scientific literature. However, there is evidence that these sources often lack racial data. As a result, the accuracy of AI tools for minority groups may be lower. This could lead to ineffective drug recommendations or racist medical advice. The researchers stated in their study paper, "It is well known that the background of descendants of minority ethnic groups is associated with differential risks in many disease populations." "If published literature is biased and inaccurate, it is logical that future AI models will perpetuate and further exacerbate these issues." Scientists are also concerned about the threats faced by low-income and middle-income countries (LMICs). AI models are primarily developed by affluent countries, which also dominate the funding for medical research. The result is that LMICs are "severely underrepresented" in medical training data. This could lead to AI tools providing incorrect advice to people in these countries. Despite these concerns, the researchers acknowledge the benefits that AI brings to the medical field. To mitigate the risks, they propose several measures. Firstly, they hope that models will clearly describe the data used in their development process. They also call for further work to address health inequalities in research, including better recruitment and recording of racial information. Training data should be sufficiently representative, while further research is needed to apply AI to marginalized groups. The researchers state that these interventions will promote fair and inclusive healthcare. Dr. Mohammad Ali from the University of Leicester said, "We must proceed with caution, recognizing that we cannot and should not block the flow of progress." Source: https://thenextweb.com/news/ai-chatgpt-healthcare-exacerbate-inequalities-ethnic-minorities

RECENT AI NEWS

RECENT AI TOOLS