José Luis: by José Luis Cárdenas Tomažič

Adapted from the Spanish column originally published in El Economista on April 21st, 2023 (Inteligencia Artificial y su Impacto en Salud)

In recent months, the social conversation around the rapid advancement of artificial intelligence (“AI”) and its multiple applications in our society has reached an unprecedented level of relevance. Healthcare is no exception.

Forms of artificial intelligence, such as deep learning algorithms, machine learning, and neural networks, are being intensively explored for new healthcare applications in areas such as medical imaging diagnostics, risk analysis, lifestyle management and monitoring, health information management, and virtual assistance. Expected benefits in these areas are substantial and include faster imaging and diagnosis, improved understanding of predictive detection, and a significant reduction in costs and inefficiencies in healthcare delivery. However, AI-based clinical tools also raise a number of situations in which commonly accepted ethical values and principles—as well as their legal implications—may be challenged. This has been the subject of a recent 2022 study by the European Parliament titled “Artificial Intelligence in Healthcare – Applications, Risks, and Ethical and Societal Impacts.”

For its part, consulting firm Accenture estimated that by 2021, hospitals would invest $6.6 billion annually in AI-related technologies, primarily in developed countries. Similarly, researchers Safavi and Kalis estimate that AI applications could generate up to $150 billion in annual savings for the U.S. healthcare system by 2026. The National Bureau of Economic Research goes even further, estimating savings of between 5% and 10% of total U.S. healthcare spending—equivalent to between $200 billion and $360 billion.

One of the most important applications of AI in healthcare is diagnosis. AI systems can analyze large volumes of medical data—including imaging—to detect diseases and conditions accurately and quickly. For example, AI can analyze CT scans to identify signs of lung cancer or examine retinal images to detect vision problems. This is particularly relevant given that some estimates indicate that diagnostic errors account for 60% of all medical errors.

In rare diseases, the use of AI for diagnosis can be especially decisive. In fact, there are between 6,000 and 8,000 known rare diseases, affecting approximately 400 million people worldwide. It is estimated that it takes an average of five years to diagnose a rare disease. As a result, patients often invest significant time, effort, and financial resources to obtain an accurate diagnosis. Some companies are already offering AI-based DNA diagnostic services for rare diseases. Since medical professionals are not specialized in all rare diseases, they tend to focus on a limited subset, which can lead to many false positives or negatives.

Another major application of AI in healthcare is treatment personalization. AI systems can analyze a patient’s medical data—including medical history and genetic profile—to determine the most effective and personalized treatment. This can help improve treatment efficacy and reduce side effects.

In addition, AI can be used to predict and prevent diseases. By analyzing vast amounts of medical data, AI systems can identify patterns and trends that may indicate a higher risk of disease in certain patients. This enables physicians to take preventive measures and provide personalized recommendations to reduce risk.

Currently, AI is already helping doctors and hospitals worldwide improve efficiency and productivity in healthcare delivery—for example, by automating administrative tasks such as appointment scheduling and medical records management—allowing healthcare professionals to focus more on patient care. AI tools can also support better-informed decision-making and improved time management.

However, it is important to note that AI cannot replace healthcare professionals. AI systems are valuable tools that can enhance healthcare delivery, but they cannot make decisions independently. Physicians and other healthcare professionals must remain responsible for clinical decision-making, patient care, and ongoing medical education and access to knowledge.

In this context, significant legal and ethical challenges arise that must be addressed to ensure responsible and safe use.

Regarding privacy and data protection, AI in healthcare processes large volumes of sensitive personal and medical data; therefore, anonymization and secure storage must be guaranteed, in compliance with applicable data protection and privacy regulations. In terms of liability and decision-making, while AI can support more informed and precise decisions, it remains necessary to determine who is legally responsible, what standard of care applies, and how decisions can be challenged or appealed in the event of errors. Another related risk is bias and discrimination in AI tools, stemming from training data and algorithm design; therefore, it is essential to establish mechanisms to detect and correct such errors in a timely manner, as well as to address risks such as hacking or similar threats.

Regulations addressing these diverse and complex issues are still in their early stages in many parts of the world and must evolve to ensure that AI in healthcare can develop appropriately—amid voices calling for a pause in its advancement.

In conclusion, AI has the potential to transform healthcare and improve patients’ quality of life. It is already rapidly advancing in developed countries and holds significant potential for developing nations as well. As discussed, there are both risks and substantial benefits, which must be carefully balanced to establish appropriate regulations—ensuring that AI becomes a tool that generates better health outcomes for patients, always in conjunction with healthcare professionals.