Joaquim Vives, PhD, Contributing Editor
Banc de Sang i Teixits
Barcelona, Spain
The field of cell and gene therapy (CGT) has witnessed a surge in innovative products that hold immense promise for treating a wide range of diseases, with CAR-T cell technology the current flagship of successful developments. However, effective translation of candidate therapies into actual marketing authorization approvals (that will ensure broad patient access) requires a concerted effort from all the stakeholders involved (including researchers, clinicians, policy makers, and technology partners). As such, the complexity of product development requires the kind of significant effort that might actually benefit from Artificial Intelligence (AI) tools. In this editorial article, we explore the opportunities and limitations of leveraging the application ChatGPT (a chatbot developed by OpenAI) to accelerate the research, development, and clinical translation of CGTs.
Opportunities of Artificial Intelligence in Cell and Gene Therapy Research
Long ago, in the 90’s, I learnt that “neural networks”, a machine learning process inspired by the human brain through a set of algorithms, could be used to improve phylogenetic classification (and discovery of new species), or even to assist with early detection of breast cancer in mammograms. A broader concept of “deep (machine) learning” was further developed to automate the learning process, therefore eliminating some human interventions. Encompassing these previous concepts, today’s “AI” has a much broader scope than those initial neural networks. AI refers to the capability of a computer system to mimic human cognitive functions such as learning and problem-solving, which are used to predict, automate, and optimize human tasks, e.g., speech and facial recognition, decision making, and translation.
In the last few months, a lot was said (especially on Twitter) about the potential uses of ChatGPT due to its ability to process vast amounts of data and synthesize information. This type of AI tools could assist researchers to advance our understanding of the biology of cell transplantation by, for instance, analyzing complex data sets, including genomic and proteomic data, facilitate the design of experiments, therefore accelerating the research process. Such processing capabilities can help researchers sift through a large volume of scientific literature and identify relevant studies. This can be particularly useful in identifying knowledge gaps, areas of consensus or controversy, and future research directions. Moreover, this technology could be used in the design of clinical trial protocols, identify patient populations that are most likely to benefit from cell and gene therapy and propose personalized treatment plans based on individual patient data.
Impressive! No wonder why, for some, ChatGPT is becoming the new member in the lab!
Limitations and Risks
Despite the tremendous potential and simplicity of use of ChatGPT, there are several limitations and risks that need to be addressed before it can be safely and effectively used in CGT development.
Firstly, biases in the training data could lead to biased predictions, which could have serious consequences for patient outcomes. Therefore, it is critical to ensure that the training data is diverse and representative of the population of interest.
Secondly, there is an associated human-factor risk of scientific rigor being compromised in the race to develop innovative therapies. ChatGPT-generated insights should be validated through rigorous experimentation before being translated into clinical practice. Researchers should also ensure that they adhere to the highest scientific standards, including the use of appropriate controls, blinding, and statistical analysis.
Thirdly, there is a risk of ethical considerations being overlooked in the pursuit of innovative therapies. Researchers should ensure that their research is conducted ethically, with appropriate consideration given to issues such as informed consent, privacy, and data security.
Finally, there might be conflicts of interest and privacy breaching with the information provided in your requests to ChatGPT. All information uploaded to this application (maybe confidential data; maybe non-anonymous data) will likely be incorporated as new data for training the algorithms.
It’s about the question
In conclusion, ChatGPT offers a potential powerful tool for accelerating the research, development, and clinical translation of novel CGTs. Beyond the potential value of the answers that this AI chatbot can produce, it is critical to consider “how” the request is formulated and “what” information included may have the key to provide the most useful insights delivered by the ChatGPT technology.
#feature4
#ISCTCommunity
#fromtheeditors#ISCTCommunity