By Wu Chen-Chi (Professor, School of Medicine, National Taiwan University Director, e-Health and Law Research Center, International Artificial Intelligence and Law Research Foundation)

Generative AI (GenAI) is a class of algorithms and machine learning models capable of generating content such as text, images, audio, video, and code. It possesses numerous advantages, including the ability to search and analyze vast amounts of data, convert between media formats and languages, create new content and dialogue, assist in information management, and boost productivity. In the healthcare sector, the application potential of Generative AI is immense, covering pharmaceuticals, medical devices, and clinical settings.

The Rapidly Evolving Landscape of AI Healthcare

In the pharmaceutical industry, Generative AI can be applied across various stages, including drug development, manufacturing, logistics and supply chain management, as well as sales and customer service. In drug development, it can analyze literature and data to discover new drug candidates, optimize clinical trial designs, and accelerate the R&D process. In manufacturing, GenAI can optimize production workflows, monitor quality, and predict supply chain conditions, thereby increasing production efficiency.

In the medical device industry, Generative AI is applicable to both R&D and sales services. Regarding R&D, it can accelerate software development processes and automatically generate documentation and records, improving design efficiency. In sales and service, it can analyze market information to enhance service quality, simplify product explanations, and improve user experience. For instance, Medtronic in the U.S. collaborated with NVIDIA to develop an AI tool for colonoscopies, used for real-time analysis of colon imagery to improve detection accuracy.

In clinical settings, Generative AI can be applied to medical practice and hospital management. It can automatically generate medical record documentation, assist in diagnostic judgment, and optimize hospital management workflows. For example, the Mayo Clinic in the U.S. partnered with Atropos Health to develop a physician consultation service platform, utilizing Generative AI to analyze Real-World Data (RWD) and provide treatment recommendations to clinicians.

Overall, the application of AI in the medical and pharmaceutical fields will significantly improve work efficiency and the quality of care. However, numerous challenges remain to be overcome, such as talent cultivation and information security. As technology continues to evolve, Generative AI will integrate with other emerging technologies to jointly push the healthcare industry toward a more efficient and human-centric new stage.

Legal Challenges in AI Healthcare

The deployment of Generative AI in the medical field will inevitably give rise to new types of legal challenges, including issues related to privacy, medical device regulation, competition law, intellectual property rights, cybersecurity, and product liability.

1. Privacy Rights

Privacy is a major concern. To ensure medical privacy, the concept of "Service on the Cloud, Data on Site"—where services are deployed in the cloud while medical data remains stored within individual medical institutions—has become a development trend in recent years. However, since Generative AI requires massive amounts of personal data for training, it may still touch upon data protection regulations in various countries. As smart healthcare becomes ubiquitous, public information privacy will require stronger guarantees.

2. Regulatory Frameworks

Smart medical devices differ from traditional ones. Smart devices are software-centric, have short product lifecycles, and require frequent updates or modifications throughout their lifespan. Regulation should adopt a more flexible model, formulating corresponding simplified listing procedures for different products, combined with post-market surveillance, to balance product safety with industrial innovation.

3. Competition Law

Scholars have proposed diverse future trends regarding market structure, ranging from an ecosystem where "a hundred schools of thought contend" to a market dominated by a few large tech giants (such as Google and Microsoft). Regardless of the outcome, issues regarding competition, privacy, and market monopolization must be monitored, with antitrust regulators playing a key role.

4. IP, Cybersecurity, and Product Liability

  • Intellectual Property: Generative AI uses vast amounts of data, potentially leading to disputes over copyright ownership and authorship. Existing regulations are difficult to apply, necessitating a new legal framework.

  • Cybersecurity: While GenAI can help improve medical network security, its own potential security vulnerabilities must also be noted.

  • Product Liability: If a diagnostic error occurs while using AI software, patients may pursue product liability claims. Regulations regarding digital products need to be strengthened to protect consumer rights.

In summary, while the medical application of AI brings opportunities, it also brings challenges. A comprehensive legal framework must be established, and supervision and enforcement strengthened, to ensure that patient privacy and safety are fully protected while promoting the sustainable development of the industry.

Back to the top