How can medical AI harm health? : Independent

Doctors have warned of increased reliance on artificial intelligence in the medical field, which may put our health at risk in various ways.

Artificial intelligence is increasingly being used in fields such as radiology, to quickly examine medical data and images. But doctors said this situation could also contribute to climate change and disastrous effects on people's health.

“According to the World Health Organization, climate change is the biggest health threat facing humanity,” say the authors of a new research paper focusing on radiation, explaining that the large amount of emissions coming from the medical field could harm people’s health.

“Medical imaging generates a lot of greenhouse gas emissions, but we often don’t think about the environmental impact of its associated data storage and AI tools,” said Kate Hanneman, vice-president of research and associate professor from the University of Toronto. “Developing and deploying AI models also consumes "Large amounts of energy and data storage needs in medical imaging and artificial intelligence are growing dramatically."

“We need to strike a balance, achieving positive impacts while minimizing negative impacts,” she added. “Improving patient outcomes is our ultimate goal, but we want to do it while using less energy and generating less waste.”

Developing intelligent models for use in healthcare requires collecting a huge amount of data, in addition to the billions of medical images that are created every year. All this data should be stored in server centers.

These centers use enormous energy, and “recent estimates indicate that total global greenhouse gas emissions from all data centers are greater than the aviation industry,” Hannemann said.

Experts suggest that AI should be designed to be efficient, use hardware that requires less power, and ensure data is compressed and removed when it is redundant.

The new paper, “Environmental sustainability and artificial intelligence in radiology: a double-edged sword,” is published in the journal Radiology.