The age of AI in healthcare: disrupting efficiency & impacting ethics

By Amit Shukla

Artificial Intelligence (AI) in healthcare is finally coming of age. Besides predicting epidemics, diagnosing illnesses and treating patients, it is now facilitating enhanced patient experience. It is also augmenting the role of doctors, by enabling them to leverage what was once an unmanageable mass of medical information hidden in legacy healthcare systems and experiential information gathered over the years by thousands of doctors.

Thanks to AI, doctors can now use this information to analyze the potential of different treatments, to find the most effective one. In fact, AI can suggest treatments even for challenging cases. It can help radiologists analyze scans and images; recommend treatments for simple cases, eliminating a visit to the doctor; and even enable remote health monitoring and medication intake tracking.

In addition, AI is helping overcome the pressing need that the industry is facing for data mining, enhanced analytics and better decision-making capability. This has placed AI at the core of the digital transformation that the industry is going through. Owing to this, healthcare has become one of the foremost beneficiaries of the information accessibility and insights that AI delivers.

Augmenting Human Efficiencies

Infosys recently commissioned a study with Vanson Bourne, to understand the opinions of key decision-makers at large organizations, about AI. It revealed that close to 50 percent respondents in healthcare have already embarked on the AI adoption journey. This places healthcare, which has just recently been through a similar overhaul of legacy systems in favor of electronic health and medical records, above others like retail, financial services and the public sector.

AI deployment is set to grow radically. With its potential to connect various stakeholders in the healthcare ecosystem, like doctors, medical device companies, pharmaceutical companies, technology providers, insurance providers and even caregivers, it will soon be among the key drivers of connected healthcare. It will enable all the stakeholders to collect, share and store information, to ultimately collaborate with each other.

AI will soon go beyond its current mandate and facilitate end-to-end healthcare for every individual – from prediction to treatment. In fact, AI’s capability to transform is not restricted to just one industry, state or country. This is perhaps why, according to the Infosys study, India is expected to be among the first to reach “mature AI adoption”, in about 2.5 years on average. Interestingly, the study reveals that India is among the top 2 countries with the highest AI maturity score.

Gaining Purposefulness with Ethical Considerations

While the capability of AI in enabling decision-making is being lauded, its ability to make decisions is being viewed with skepticism. Many in the industry are wondering if machines should be allowed to take decisions, albeit eventually. I believe that as long as the algorithms that make these decisions rely only on the symptoms and clinical history shared with it, we are in virtually safe hands. But, if AI systems take decisions with non-curated information, then it could create problems.

Another valid concern that many skeptics have raised is, what if the algorithms see a pattern where none exist or acquire a bias because of partial or skewed data? Then, would its diagnosis be reliable? Also, while it’s true that AI systems can predict negative healthcare outcomes – like a heart attack – with relatively high accuracy, should we really share that information with the patient? How do we discern what information is passed on and what is not?

And, even though AI leverages millions of data sets to make a diagnosis, a doctor’s diagnose will be based on years of experience and intuition. Should we replace this? These are questions to which we have to find answers, eventually. Till then, we wait and observe the progress AI makes. In the Infosys survey, it was revealed that close to 53 percent respondents from the pharmaceutical and life science industries think that they have completely considered AI’s ethical implications. This is critical.

Striking a Balance Strategically

It is essential to address the ethical concerns that AI brings along with it, so that sometime in the near future, companies can leverage its full potential. In fact, a good starting point can be – setting the boundaries for what you allow AI to do and what you limit it from. I think the key principle is to start using AI in clinical settings to assist and co-work with humans, who will be the ultimate decision makers in diagnosis and treatment.

We will only stand to benefit from having an open mind to enabling healthcare decision makers to collaborate with AI that humans can selectively depend on for information to enhance care delivery. As an early technology adopter, the healthcare industry can leverage AI to unveil its potential, dispel the prejudices around it and build awareness among professionals. To achieve this, the industry will have to set ethical standards for organizations and metrics to measure AI system performance.

In fact, the professionals whose roles are being augmented by AI can be retrained to take over the ethics charge. Lastly, the system should be allowed sufficient time – based on factors like organization maturity and technology complexity – to detect any challenges in the AI system. These steps can go a long way in helping AI realize its full potential in the industry and empowering people to leverage it for valuable assistance.

The author is associate vice president – healthcare, Infosys.