Healthcare Trends
ChatGPT: Blessing or curse?
When AI specialists talk about future technologies, the average Joe Bloggs is not likely to be all that interested. However, when even the nightly news is reporting on ChatGPT – the current star of AI – that means the topic is relevant for all of us, including HCPs.
It is well worth starting with a short explanation of the abbreviation ‘ChatGPT’. GPT stands for ‘generative pre-trained transformer’. While this term may sound quite abstract, it becomes clearer when you look at the function of the technology. A generative pretrained transformer can communicate with people and answer questions, which is why the word ‘chat’ is added at the start.
For us, this prompts the question of whether the healthcare industry might be threatened by this new artificial intelligence (AI). Can ChatGPT maybe even help manage digitalisation better, reduce the burdens that people face and drive progress in the healthcare market?
Schools and universities are taking a critical view of the new program as they fear for their students’ own performance, while other organisations are dreaming of the full automation of their customer service. This is because ChatGPT delivers what seems to be a compelling answer to any question, and that is what the AI was trained to do by using vast amounts of text data from the internet. The outstanding performance in natural speech processing tasks makes ChatGPT ideal for dialogue system automation – in customer service, in education and in fields in the healthcare industry.
What concrete applications are conceivable?
ChatGPT offers a variety of possibilities for healthcare communication, for example, improved customer loyalty or reduced workloads for teams in customer support, sales and marketing. The AI text generator can help provide fast and personalised answers to enquiries in customer support departments without human employees needing to be available constantly. This lets companies save on costs whilst simultaneously improving the customer experience. It can also be used to gather feedback, of course, and obtain new findings about products or treatments based on it.Personalised medicine recommendations are also an idea since ChatGPT can suggest a suitable medicine to the treatment provider using the patient data. The program can analyse large volumes of data and create prediction models, so companies will in the future have a better ability to develop new medicines and optimise treatments. The prediction models can also help identify the risks or benefits of a medicine early on.
ChatGPT even itself claims to be able to generate molecules that could potentially serve as new active ingredients. Using machine learning and deep learning techniques, the new AI is supposed to be able to analyse the suitability of existing molecules as active ingredients. Furthermore, the program can be used when training medical personnel to expand their knowledge and abilities. For example, training sessions can be offered in a more efficient and cost-effective way by using chatbots.
Even with all these positive aspects, however, ChatGPT also represents a challenge – especially when the priority is on how binding the information is, on complying with rules and regulations and on safeguarding legislative requirements. The sources of data on which this AI text generator is based are not always verified, so how correct is the information? For as long as brands cannot ensure that the AI is trained solely on verified data sets that they provide, it is borderline impossible to use ChatGPT as a standard tool.
The use of ChatGPT also presents risks in relation to the protection of patient data privacy. Regardless of each user needing to give their consent to the processing of their data, there is no doubt that uncontrolled usage should not be allowed. There is good reason why the usage of medical data is so strongly regulated. To uphold these high standards, a comprehensive code of ethics is indispensable for the use of chatbots and other AI-controlled solutions. It is absolutely essential to comply with the regulations issued by the FDA, EMA and other supervisory authorities. This also includes the requirement that sources of data supporting chatbot answers be transparent and limited to verified sources by their owner, who in turn must hold liability for them.
In summary, ChatGPT can be described as an additional tool for experts in the healthcare sector, though not a replacement for them. Rather, this AI demands professionals who know how to use it correctly. Chatbots need the support of experts because they cannot make miracles happen by themselves. Accordingly, the greatest challenge for pharmaceutical companies is finding talented people who are able to use AI tools effectively and simultaneously ensure adherence to legal requirements and the accuracy of the information.