The liabilities for AI in healthcare

The healthcare industry has benefited greatly from the technological advances of artificial intelligence, from deep learning algorithms that can predict and diagnose diseases to natural language processing that can process voluminous amounts of unstructured health records and drug safety data, the potentials for AI in healthcare seem endless. However, every coin has two sides, and AI is no exception.

Data privacy and security

AI technology is increasingly used in the healthcare sector to process massive datasets. Such datasets will inevitably include personal health information, generally regarded as sensitive personal information requiring a higher degree of protection. Fortunately, general public awareness of privacy and data protection issues is increasing. As the use of personal information becomes more regulated worldwide, stakeholders must give careful thought to their data handling and security practices, lest they incur the financial and reputational repercussions of a data breach.

Liability

When a doctor makes a mistake, the doctor may be subject to medical malpractice lawsuits. However, when an AI-powered equipment makes a mistake, who should be held accountable?

Most applications of AI in the healthcare industry require input from physicians who serve as the main drivers of decision-making. As AI continues to evolve and becomes more independent in performing tasks, or as physicians become more reliant on the recommendations made by AI, there will be more legal scrutiny on physicians and healthcare institutions that incorporate such AI technologies in their practices. Where there is use of AI, there is potential for liability.

In addition, as AI technologies continue to advance, it is becoming increasingly difficult for humans to dissect their decision-making processes. This may be particularly true for technologies based on deep learning, where the black-box nature of their processes makes it difficult to determine the rationale behind a decision. In such instances, assigning responsibility to an individual may be difficult when something goes wrong.

The gray zone

There is currently limited legal precedent regarding liability involving AI. Should a physician or healthcare institution be liable for errors or malfunctions of AI technology, or should the developer or vendor of the AI technology be responsible for its errors?

Privacy regulators in Canada recognize the benefits of AI and the liability and privacy challenges that AI adoption raises. Several projects funded by Canadian privacy regulators have been completed, and more will likely be under way to help regulators and legislators understand how best to address the liability and privacy risks arising from the application of AI in healthcare and other industries.

As AI continues to develop and its applications in various industries expand, novel legal questions will arise. To maximize the benefits of AI technologies, it will be important for industry leaders and regulators to continually monitor the development and the impact of AI technologies in their industries, identify the pros and cons that arise from their adoption, and guide other industry professionals and the public regarding their appropriate uses and applications.