The introduction of artificial intelligence (AI) has revolutionized healthcare systems. A new paper in the journal Diagnostic and Interventional Imaging discusses the steps necessary to ensure AI in healthcare is used responsibly and sustainably.
Study: Climate change and artificial intelligence in healthcare: Review and recommendations towards a sustainable future. Image Credit: metamorworks / Shutterstock
Background
AI has been incorporated into healthcare devices, mostly in diagnostic imaging, radiation therapy, interventional radiology, and nuclear medicine. Deep learning (DL) is the most frequently used AI application in such devices. DL allows models to learn from data without the operator's involvement and thus can improve diagnostic and therapeutic outcomes and increase care efficiency.
However, this computationally heavy platform comes with a high carbon footprint, which may accelerate climate change and negatively impact the environment in multiple ways.
Climate change is an urgent issue, and the need to mitigate its effects and slow its rate of progression has been recognized internationally. The current review assessed the use of AI in healthcare in the context of climate change.
The pros of AI for healthcare and climate change
The positives of AI in healthcare include a much smoother, faster, and less wasteful workflow and the ability to use telemedicine more widely. AI can reduce the waste of resources such as energy, time, and imaging materials by improving the identification of patients requiring imaging and reducing wait times.
AI can also enhance the clinician’s diagnostic capabilities, thus avoiding the need for repeat examinations. When coupled with a sleeker workflow, this can help promote virtual care and reduce unnecessary patient travel, bringing down greenhouse gas emissions.
The energy cost of AI in healthcare
The training and utilization of AI models are energy-intensive. Healthcare applications account for over 4% of AI use at present, and they require large datasets, complex algorithms, and multiple model updates. One study reported that a single large AI model takes as much energy to run as five cars over their entire lifespan.
The use of AI in healthcare depends on data centers that use servers, cooling systems, and networking platforms. All these must run constantly in controlled environments, consuming a lot of energy and accounting for about 1% of global power consumption.
Healthcare also produces large amounts of electronic waste due to the constant need for hardware updates. Such waste can poison the environment due to the use of materials such as lead, cadmium, and mercury.
The high demand for natural resources like rare earth elements takes a toll on biodiversity by promoting habitat destruction. Transportation and AI-associated supply chain logistical demands intensify the indirect impact of healthcare-linked AI on the environment.
Mitigation measures
Possible solutions could include increasing the energy efficiency of AI models via techniques like quantization and pruning. Improved infrastructure design, revamping hardware and software concepts, and efficient power management using dynamic voltage and frequency scaling can also reduce AI's environmental costs.
Incorporating renewable energy can reduce AI-associated energy consumption. In fact, AI-aided nuclear fusion reactor design could make progress in harnessing this power source for AI in healthcare.
Such steps require a comprehensive lifecycle evaluation for environmental cost, enabling scientists to seize opportunities to reduce the carbon footprint from beginning to end. One study reported that “autonomous AI could potentially reduce greenhouse gas emissions in healthcare by up to 80%”.
Cooperation of stakeholders is key
Sustainability practices for AI in healthcare will only be successful if policies and government initiatives are strengthened. This requires collaborating with stakeholders at all stages of the process.
Regional and international cooperation is essential for these trends to become the norm, and the sharing of knowledge is essential.
Best practices for sustainable AI in healthcare include designing green frameworks, AI system lifecycle assessment, the responsible use of data, and regulatory oversight of changes and movements in the field. By understanding the current state of knowledge, this review of AI in healthcare and its impact on climate change aims to direct future research and target areas where better practices are required.
Prioritizing sustainability and environmental responsibility is crucial to ensure that the benefits of AI are realized while actively contributing to the preservation of our planet.
Review highlights advances in flexible optoelectronics for cardiac healthcare