GPT-4: The Future of Human-Centered AI in Healthcare

A design philosophy that begins with empathy for users and pushes forward with humility about the limits of machines and people will help build more reliable, safe, and trustworthy systems. “Empathy enables designers to be sensitive to the confusion and frustration that users might have when AI systems fail. Humility leads designers to recognise the need for audit trails that can be retrospectively analysed when the inevitable failures occur.” — Human-Centered AI, Shneiderman, B. 2022. 

Since its rollout in November 2022, a host of ethical issues has arisen from the use of ChatGPT: bias, privacy, misinformation, and job displacement, among others. But this did not dampen public enthusiasm. By February 2023, ChatGPT broke the record for application user-base growth, rising to 100M in just two months (Last Week in AI, 2023) and reaching a market value close to 30 billion USD (Roose, 2023c). 

In 2022, venture capitalists invested 2.7 billion USD in 110 generative-AI startups (The Economist, 2023a). In consequence, alternatives to ChatGPT did not take long in appearing: Deep Mind’s Sparrow, Google’s LaMDA/Bard, Antrophic’s Claude, and an open-source transformer by Stability AI (Konrad and Cai, 2023b; Metz and Grant, 2023; Romero, 2023f; Roose, 2023c,d).

Our world is grappling with many challenges to global health. With an increasing incidence of pandemics, a surge in chronic diseases, mental health concerns and a rapidly ageing population, the healthcare industry is undoubtedly struggling to keep up with the rising demand for high-quality healthcare services.

According to the World Health Organization (WHO), there is a 1 in a million chance of a person being injured while travelling by aircraft. In comparison, there is a 1 in 300 chance of an individual being harmed throughout the patient journey. Prior research has shown that up to 50% of all medical errors in primary care are due to administrative reasons. The global shortage of medical professionals further compounds these problems. The WHO also estimates a projected shortfall of 10 million health workers by 2030, mostly in low- and lower-middle-income countries making it increasingly challenging to provide care to everyone in need.

AI and predictive analytics help us to understand more about the different factors in our lives that influence our health, not just when we might get the flu or what medical conditions we’ve inherited, but things relating to where we are born, what we eat, where we work, what our local air pollution levels are or whether we have access to safe housing and a stable income. These are some of the factors that the World Health Organization calls “the social determinants of health” (SDOH).

In 2030, this means that healthcare systems can anticipate when a person is at risk of developing a chronic disease, for example, and suggest preventative measures before they get worse. This development has been so successful that rates of diabetes, congestive heart failure and COPD (chronic obstructive heart disease), which are all strongly influenced by SDOH, are finally on the decline.

In recent years, we have seen artificial intelligence (AI) transform from a concept of science fiction into a reality that touches multiple facets of our everyday lives. AI has rapidly evolved, and its growing influence on various sectors is undeniable. Among these, the healthcare sector stands as a testament to the transformative potential of AI.

One of the most impressive exemplars of this technological revolution is OpenAI's Generalized Pre-trained Transformer 4 (GPT-4). This advanced language model leverages machine learning algorithms to predict the next word in a sentence based on the context. The sophistication of GPT-4 goes far beyond this rudimentary description. The model’s scale, measured in terms of parameters, allows for highly accurate, contextually relevant responses.

The healthcare sector, with its relentless flow of data and constant demand for precision, can benefit enormously from GPT-4. It can revolutionise information management by processing, analysing, and extracting crucial insights from vast quantities of data, presenting it in an easily digestible format. Such capabilities could enhance the efficiency of healthcare services, ultimately benefiting the patients.

Telehealth, a rapidly growing field in the wake of global health crises, also provides fertile ground for the application of GPT-4. Here, this AI model can serve as a reliable first point of contact for patients, helping to answer health-related queries, suggest possible diagnoses, and refer to appropriate medical services, all while reducing the strain on healthcare professionals.

Additionally, GPT-4 has the potential to become an instrumental tool in promoting health education and patient awareness. It can generate comprehensible content from complex medical information, thereby improving health literacy, which is crucial for empowering patients in managing their health.

However, while embracing the impressive capabilities of GPT-4, it is crucial not to lose sight of the importance of a human-centered approach. Despite the incredible strides in AI, it is a tool that augments human capabilities and should not be seen as a replacement. Empathy, compassion, nuanced judgment, and the human touch remain at the core of healthcare, elements that no AI, no matter how advanced, can replicate.

As we stand on the precipice of this technological revolution in healthcare, we must ensure that our strategies for AI integration emphasise human-AI collaboration to provide the highest standard of care. AI can assist and even revolutionise many aspects of healthcare, but it is the humans behind the machines that make a difference in patients lives.

The potential of GPT-4 and its future role in healthcare is both exciting and vast. However, its successful implementation will depend on our ability to synergise the efficiency and precision of AI with the irreplaceable empathy and understanding of human caregivers. With a balanced approach, we can ensure that the future of healthcare is not just technologically advanced but also humanly compassionate.

There is a genuine gap in addressing the importance of Human-Centered AI (HCAI) that goes beyond the perspective of engineers alone. While engineers play a crucial role in designing and developing AI systems, addressing the broader implications and considerations of HCAI requires input and perspectives from diverse stakeholders, including ethicists, social scientists, policymakers, and end-users. The discussion surrounding HCAI involves not only technical aspects but also ethical, societal, and human-centered dimensions. It requires interdisciplinary collaboration and engagement with various stakeholders to ensure that AI systems are designed and deployed in a manner that prioritises human values, addresses societal concerns, and respects individual rights.

Non-human actors will become part of our society in the near future, and everyone will inevitably be using non-human social actors (e.g. Smart IoT Devices, AI-Powered Wearables (Medical Monitors, Preventive Care Devices, Access Control, Smart Clothing, Automobiles, et al.), Nest Thermostats, iRobot Vacuum Cleaner, Alexa, Smart Appliances) in their own homes within the not-so-distant future. Design, societal, legal and ethical issues cannot be addressed without empirical data on the relationship between humans and non-humans. There is a need to involve users early in the design process to understand the consequences of such interactions on an individual and societal level. By understanding people's acceptance of autonomous systems, researchers can more effectively address their research efforts toward designing responsible AI. 

AI technology has brought in a paradigmatic change in the human-machine relationship. HCI work has traditionally focused on the interaction between humans and non-AI computing systems. These systems rely on fixed logic rules and algorithms to respond to the input from humans; humans interact with these systems in the form of “stimulus-response” (Farooq, U., & Grudin, J. (2016). These computing systems (e.g., automated machines) primarily work as assistive tools, supporting human monitoring and execution of tasks (Wickens et al., 2015). As we are currently entering the AI-based “autonomous world,” AI technology has given new roles to machines in human-machine systems as driven by their autonomous characteristics.