Since the beginning of 2023, ChatGPT has been on everyone’s lips. By now almost everyone knows broadly how it works, but perhaps not everyone is aware of the fact that it is part of the large family of Generative AI, a branch of Artificial Intelligence that involves learning a representation of data to generate new, original artefacts, such as texts and images.
Beyond the hype generated by ChatGPT, it is worth noticing that the landscape of Generative AIs is extremely vast, and their future impact may be transformational to various industries, changing the way we work every day.
Going back to ChatGPT, it is a perfect mix of emergent abilities and areas of improvement. Whether it masters general-purpose knowledge, admits its own mistakes, challenges incorrect premises and remembers past interactions, ChatGPT is also exposed to the risk of hallucinations, it is only updated until 2021, and it is not able to handle sensitive data and guarantee data privacy and security. These aspects, both pros and cons, are of particular relevance in the pharma and life sciences context.
ChatGPT Use Cases: Real-World Applications in pharma & life sciences
In order to disentangle the potential of ChatGPT in the pharma and life sciences world, we have analysed 6 use cases, 3 addressed to pharma companies and the other 3 directed to healthcare professionals (HCPs) and patients. These applications currently lie in different stages of readiness, varying from use cases that can be replicated with the simple use of ChatGPT, to instances requiring APIs and integrations.
For pharma companies we thought that ChatGPT could be used:
- To enhance customer service in the pharma industry by providing real-time responses to customer queries and improving the overall customer experience, although still requiring close supervision;
- Even more effectively, for the creation of content such as product descriptions, drug labels, and marketing materials for the pharma industry;
- To create synthetic data that can help researchers and pharmaceutical companies on several applications, without risking patient privacy.
On the healthcare professionals’ and patients‘ side, ChatGPT could be used:
- To assist HCPs in navigating medical literature to find and synthesize relevant information from a vast and complex collection of medical literature;
- To create a virtual symptoms tracker helping patients monitor their health status and providing insights on potential conditions;
- To act as a virtual assistant to HCPs with tasks such as scheduling appointments, managing patient data, and answering common questions.
ChatGPT (real) applications: what does the future hold?
Although it has already generated significant hype, ChatGPT is just the seed of what promises to be a revolution. Many huge tech companies have started working on similar technologies, such as Google with Bard. In the meantime, OpenAI is working on GPT4, a more powerful and capable language model than GPT3. New careers and roles built precisely on the ChatGPT model are emerging, such as the one of Prompt Engineer, and there is a wide spectrum of sectors already embedding this technology in their operations, ranging from fashion and well-being to tourism and jurisprudence.
Net of all these current and potential applications, it is crucial to approach the ChatGPT issue in the right, balanced way. At this stage, the common sentiment oscillates between positions of extreme excitement and deep disillusionment, following a sort of Dunning-Kruger effect when instead the phenomenon should be observed in its wholeness.
ChatGPT can represent an extremely useful tool that can provide quick and accurate responses to a wide range of inquiries and significantly enhance productivity. Nevertheless, it is essential to recognize its limitations in terms of output quality, data privacy and security, and technical complexity.
What about pharma companies?
Given the sensitivity and complexity of the field, the recognition of both ChatGPT’s capabilities and limitations is essential to address this issue in the proper way: with the exception of pure content creation, its immediate implementation may not be possible due to the need for development and integration work, and some form of supervision is unavoidable anyhow.
Therefore, a pharma company should clearly determine the use cases of interest, conduct risk assessments, develop an integration plan, and effectively implement the solutions with constant and structured monitoring of the processes. To do so, it is essential to choose a reliable vendor able to provide, integrate, and support the implementation of generative AI technologies within the company.
