Machine learning and artificial intelligence (AI) is making rapid strides in the healthcare sector in India. Global technology leaders such as Microsoft, Google and Siemens as well as dozens of new Indian startups are introducing promising AI solutions to address the country’s burgeoning healthcare needs.

Push for Faster Inclusion of AI

With India’s untenable patient-doctor ratio, global analysts expect that the pressure to improve healthcare delivery will lead to a faster inclusion of AI and Deep Learning (DL) into the country’s existing diagnostic and treatment processes.[1]

According to a 2017 Accenture report, by 2035, AI could add nearly US$1 trillion or 15 percent of current GVA (gross value added) to India’s economy.[2] In Q3 of 2017, as many as 16 medical IT startups in India received global funding.[3]

At the state level, the government of Karnataka plans to mobilize Rs. 2,000 crores to promote AI startups by 2020.[4] Key national stakeholders such as the PMO[5] and FICCI[6] are also pushing for the faster integration of Indian healthcare with AI.

Innovative Medical AI Solutions in India

Both large companies and startups are working to gain an early lead in India’s developing medical AI solutions market. Siemens Healthineers is offering a range of AI products that are designed to standardize and automate complex diagnostics and meet patient needs.[7]

Microsoft has announced a partnership with Apollo Hospitals to build an AI-centric cardiology network. Microsoft is also expanding its existing AI network for eye care in India. The company’s VP for AI & Research, Dr. Peter Lee, said that Microsoft will use AI models for predicting heart disease risks in patients and support physicians with targeted treatment plans.[8]

Bengaluru-based startup Niramai, which is partly funded by Flipkart co-founder Binny Bansal, is using AI to create accurate, painless and affordable breast cancer screening tests.

Max Healthcare, a leading hospital chain in North India, is deploying AI to monitor critical care. According to the company, the technology has already brought down critical care costs by about 30 percent along with optimized use of ICU beds.[9]

Pune-based LiveHealth, which recently raised $1.1 million in seed funding, is deploying AI to process millions of ERP transactions and medical records, and now plans to develop products that will assist patients and doctors to make informed decisions.[10]

AI diagnostics startup SigTuple is reducing the burden of pathologists with intelligent digital analysis of blood samples. According to SigTuple’s CEO Rohit Pandey, the algorithmic analysis by its AI bots is as accurate as that of an actual pathologist examining a tissue sample under a microscope.[11]

Bengaluru-based startup for AI cervical cancer screening, AIndra, is applying artificial intelligence to screen samples and identify the ones that indicate cancer risk. Another Bengaluru- and Singapore-based 3D printing startup Supercraft is developing AI visualization tools that equip doctors, hospitals and researchers with deeper insights into human anatomy.[12]

According to Chander Shekhar Sibal, VP and Head of Medical Division, Fujifilm India, AI can make a vital difference to the early diagnosis and prevention of diabetic retinopathy, which is currently a serious medical challenge. Sibal further says that AI can dramatically improve speed and accuracy in the area of radiology in India for early detection and containment of diseases, including malaria and HIV.[13]

Absence of a Legal Framework and Lack of Vision
While advancements in AI are India’s best bet yet to sustain its crumbling healthcare infrastructure, visionless implementation of AI medical technology and the absence of a robust legal framework can only compound the crisis, and not mitigate it.

In 2017, the Ministry of Commerce and Industry set up a high-powered “Task Force on Artificial Intelligence” to explore how AI could be leveraged for India’s economic development across different sectors.[14]

Although the 18-member task force consists of outstanding civil servants, entrepreneurs, technocrats and researchers from the fields of engineering, IP, robotics, and artificial intelligence, it does not include a medico-legal expert to address the issues of law that are likely to arise with the integration of healthcare and AI.

There is an urgent need for the policy makers, healthcare AI developers, and medical practitioners in India to consider the legal aspects while designing, implementing, and regulating artificial intelligence in the healthcare sector.

Unanswered Legal Questions
How do AI programs ensure patient consent and privacy of sensitive medical data? How to address the questions of apportionment of liability among the practitioner, hospital, and the AI system developer, trainer and manager in case of an act of medical negligence?

In the event of an AI diagnostic error, or inaccurate use of data, or a technological malfunction, who would be held liable: the practitioner, the AI developer, the specific program engineer who designed it or the AI robot (if the robot used its ‘intelligence’ to make its own decision)?

How to determine the degree of accountability of the operating physician when a wrong diagnosis or treatment occurs due to an error in the primary data feed or an AI systemic glitch?

How to determine whether the operating physician could have anticipated or recognized the flaw in the AI at some point during the treatment or surgery, and stopped the operation or switched over to conventional techniques?

AI and the Limitations of Existing Legal Concepts

The tort system in the medical arena is founded on the concept that the doctor is the trusted specialist or the expert. It presumes that the treating doctors are fully responsible for their own decisions, and consequentially, they are also responsible if the medical care they provide is proved to be negligent. But who should be held liable when the physician delivers a wrong treatment at the recommendation of the AI diagnostic tool?[15]

A recently published paper by researchers at NYU argues that medical AI (like IBM’s Watson for cancer care) should be accorded a special legal status equivalent to personhood in order to account for its current and future role in the process of medical decision making.[16] (Will this also mean setting up of separate tribunals to deal with lawsuits against AI medical robots?)

One of the issues is that the traditional tort law concept of ‘foreseeability’ may not work when AI systems perform a medical diagnosis and treatment. For an individual to be held liable for negligence, the damage that occurred must be ordinarily ‘foreseeable.’ However, AI or machine learning systems are supposed to learn from the past data and patterns and may behave in ways that the AI developers and designers may not be able to foresee reasonably.

The labour laws in India have largely remained stagnant from the days of the socialist era that began after India’s independence. Some would argue that even IT engineers come within the purview of “workmen” as defined in the Industrial Disputes Act, 1947. In order to make transition to an AI-enabled workforce in highly specialized fields such as healthcare, some of the archaic labour laws may have to be amended.

The limitation of liability as described in the IT Act, 2000 may also be unfit to operate in the era of artificial intelligence. Section 79 of the Act suggests that intermediary service providers in the field of information technology are merely the carriers of content. Barring exceptions, under Section 79, they would not be held liable for the substance of the content. This rule may have to be re-examined with the implementation of AI systems that are devised by the carriers.[17]

Key Takeaway
A basic tenet of decision-making in the field of medicine is whether the benefits of a treatment approach outweigh its risks. Instead of fearing the integration of AI and healthcare, or going to the other extreme of allowing for its blanket implementation without a well-defined legal framework, a balanced approach would be to determine where AI should be applied and what should best be left done the conventional way.

A comprehensive legal framework of checks and balances must be developed to ensure compliance so that the benefits of medical AI outweigh the risks within the context of Indian healthcare. It is time for the Indian lawmakers and the judiciary to step in and pave the way for a successful transition of India’s healthcare sector into the age of artificial intelligence and solve India’s complex healthcare problems.

End-Notes
[1]Artificial Intelligence (AI), Healthcare and Regulatory Compliance, Forbes (March 20, 2018)
[2]Rewire for Growth: Accelerating India’s Economic Growth with Artificial Intelligence, Accenture (December 20, 2017)
[3]Healthcare IT companies raise $5.5 billion in VC funding in Jan-Sep 2017, The Economic Times (October 23, 2017)
[4]Karnataka Government to Invest $6 mn in AI, Data Science Hub, VC Circle (October 3, 2017)
[5]Israeli Firms Join Hands to Deploy Artificial Intelligence, The Times of India (July 7, 2017)
[6]Re-engineering Indian Health Care, Ernst & Young (September 1, 2016)
[7]Artificial Intelligence – Siemens India
[8]Microsoft and Apollo Hospitals to Use Artificial Intelligence for Early Detection of Cardiac Diseases, Microsoft (March 8, 2018)
[9]Artificial Intelligence Lifeline for India’s Flailing Healthcare, The Financial Express (March 24, 2018)
[10]AI to the Rescue: 6 Startups that are Redefining Healthcare Market in India, Business Standard (February 26, 2018)
[11]Bots, bytes and big data: Could AI Transform Indian Healthcare, The Hindustan Times (March 5, 2018)
[12]How Innovations in AI, Virtual Reality are Advancing Healthcare in India to New Frontiers, The Economic Times (June 9, 2017)
[13]Emergence of Artificial Intelligence in Healthcare, Franchise India, (January 3, 2018)
[14]Artificial Intelligence Task Force
[15]Artificial Intelligence, Medical Malpractice, and the End of Defensive Medicine, Harvard (January 26, 2017)
[16]Hey Watson, Can I Sue You for Malpractice? Asia-Pacific Journal of Health Law, Policy and Ethics, SSRN (November 28, 2017)
[17]Indian Law is yet to Transition into the Age of Artificial Intelligence, The Wire (September 26, 2016)