Vijay Pande, PhD: Pioneering the Convergence of AI and Biology to Transform Healthcare and Drug Discovery

From the 1970s to the present day, computer technology has advanced rapidly in performance. Today, computers at the same price point are approximately 1,000 times more powerful than those a decade ago and about one million times more powerful than those two decades ago. This has enabled the software industry to sweep across the globe.

The power of technology lies in its adherence to Moore’s Law—the tech industry’s decades-long ability to exponentially reduce costs while enhancing capabilities.

In contrast, drug design and healthcare services follow Eroom’s Law, which is derived as the inverse of “Moore’s Law.”

For decades, costs in these sectors have continued to rise, reaching such high levels that healthcare expenditures now account for approximately one-quarter of the U.S. GDP (and are still increasing). Factors such as rising labor costs, clinical trial expenses, administrative overhead, and increasingly strained doctor-patient relationships have led to an aberrant growth trajectory in treatment and medical costs.

Given that Moore’s Law drives an exponential decrease in costs, while Reverse Moore’s Law leads to an exponential increase, it is evident that efforts must be made to shift the trajectory from Reverse Moore’s Law toward Moore’s Law. In other words, human-driven services (such as caregiving) must be transformed into computer-driven ones (commoditizing services through technology). This is precisely the role of artificial intelligence.

This transformation began with simple, single-use models (commonly referred to as machine learning) designed to perform error-tolerant, straightforward tasks; for example, Netflix uses artificial intelligence to recommend shows.

As artificial intelligence becomes increasingly sophisticated, its potential applications continue to expand. Generative AI methods can now produce text and images and perform complex tasks, although errors may occasionally occur.

Advances in artificial intelligence will further promote the development of AI-driven life sciences and healthcare, significantly expanding the scale of the skilled workforce or enhancing the capabilities of the low-skilled workforce. For example, AI can propose answers or suggest lines of reasoning, enabling humans to select the optimal solution, filter results, and bypass incorrect answers. This approach allows AI to be seamlessly integrated into existing workflows.

As time progresses, the proportion of human labor is decreasing, and even in fields requiring human experts—specifically those where minor errors can lead to catastrophic consequences—the trend is gradually moving toward full automation. However, this process may not completely eliminate human involvement, particularly in areas where error is unacceptable, such as diagnosis, medication prescribing, and medical procedures.

Developing artificial intelligence capable of successfully performing these specialized tasks is a key direction for the future development of AI, and it represents the natural pathway through which future advances in AI will ultimately exert the greatest impact on life sciences and healthcare.

The tremendous advances in artificial intelligence are only part of the story. As AI continues to mature, life sciences and healthcare are also undergoing transformation. Both industries are largely driven by engineering, which holds the power to revolutionize the diagnosis, treatment, and management of diseases.

In the field of life sciences, advances in gene editing, cell biology, stem cells, and robotic experimentation have enabled scientists to manipulate biological systems in entirely new ways. These advancements have not only ushered in a period of large-scale development in biology but also forged close ties with artificial intelligence (AI). Furthermore, as AI is increasingly applied in life science experiments, a powerful feedback loop has emerged: experiments enhance AI’s predictive capabilities, while AI, in turn, further refines experimental processes.

All these advances are underpinned by massive computational power and data storage capabilities, which have only recently become feasible.

In short: We have the opportunity to leverage artificial intelligence to address the most significant challenges facing healthcare and drug design.

First is the issue of healthcare costs. The exponential growth in costs is partly driven by the demand for highly skilled professionals (PhDs, MDs, nurses, etc.)—especially given that the cost of skilled labor is rising far faster than inflation. As artificial intelligence becomes increasingly capable of performing as a technical expert, it can augment the capabilities of existing healthcare personnel, enabling the delivery of medical services at lower costs. If AI incorporates emotional intelligence, it can also communicate with patients to some extent, monitor their adherence to medical advice, and alleviate the workload pressure on clinicians.

Secondly, cost reduction will enhance the accessibility and quality of healthcare services. As artificial intelligence is increasingly applied in healthcare, medical services will become more accessible to all, providing optimal care for everyone and enabling patients to receive accurate diagnoses and treatment plans at an earlier stage.

Furthermore, artificial intelligence is playing a positive role in reducing costs and improving outcomes in the development of new therapies. Just as calculus serves a foundational role in physics, AI is a key driver in unraveling the complexities of biology—complexities that exceed the limits of human understanding. Today’s AI models can propose more effective therapeutic regimens, accelerating drug development while reducing the likelihood of failure. In short, AI surpasses human scientists’ capacity to comprehend biological systems, enabling research at a scale far beyond current paradigms, which rely primarily on serendipitous discoveries arising from hours of manual labor in the laboratory.

However, the potential issues associated with artificial intelligence are equally significant. Training early AI models on human-collected data may introduce embedded algorithmic biases or other errors. As AI is applied to new industries, scientists and regulators must remain vigilant about potential side effects.

In fact, within the life sciences and healthcare sectors, the current regulatory framework mandates testing for the efficacy and adverse reactions of all devices. Concerns that artificial intelligence (AI) is a “black box” are unfounded, as AI systems can be fully audited. Given sufficient time, any AI model can be understood in detail. Ironically, in healthcare, it is human reasoning that truly constitutes a “black box.”

Obviously, this transformation cannot happen overnight, as healthcare (and biopharmaceuticals) is essentially a conglomerate of multiple interwoven industries subject to regulatory oversight. In the short term, the outcomes brought about by artificial intelligence are likely to be disappointing. This gradual transitional process often fails to meet people’s expectations. However, in the long run, this situation may change.

To address the most significant challenges facing the healthcare and life sciences sectors, there is a need for domain-specific specialized artificial intelligence, driven by an AI company built around dedicated teams and purpose-built large language models. To this end, we believe that builders must understand the following points:

a) How to leverage the latest and most powerful artificial intelligence technologies;

b) How to commercialize products or platforms in the biopharmaceutical and healthcare sectors with reliable product and go-to-market strategies.

In this new era, teams with deep expertise in both of the aforementioned areas—comprising scientists, AI specialists, healthcare builders and operators, and product and go-to-market experts—will be most likely to lead and prevail.

Through artificial intelligence, every nurse can become a superhero for inpatients. Meanwhile, this has sparked industry reflection: What does it mean if every patient can have an on-demand, professionally trained caregiver for just a few cents? In terms of treatment, therapies aimed at extending healthspan are garnering significant attention, and AI can surpass human scientists in developing superior antibody therapies to combat human diseases.