by Diana Ribeiro – Freelance Medical Writing Consultant
Whether you are an early adopter, an unhurried user, or even a doubter, the fact is that artificial intelligence (AI) has taken the world by storm, including life sciences and healthcare. AI is not new, but it seems to have reached an inflection point. In our previous article, we explored the different facets of AI, explaining terms like machine learning, robotic process automation, and natural language processing. In this article, we will explore what AI tools can do for different areas inside the life sciences field, but also what these tools cannot (yet?) do. This will enable you to use AI tools with confidence and in an ethical way.
Artificial Intelligence Life Sciences: Summarise and Analyse
A few decades ago, the average person would have less information at their disposal than the time available it would take to read it. These days, we have more information at our fingertips than we would ever be able to read in a lifetime, let alone digest it and make connections between information sources. We are still able to do this, of course, but at a much smaller scale than AI-powered natural language processing can.
This is why different areas like medical affairs, medical writing, and scientific departments can greatly benefit from natural language processing tools that can extract useful information from scientific and medical research papers. By using these tools to summarise the main points of several texts, professionals in these areas can have more time to make decisions and draw conclusions, without spending time analysing data manually.
In addition, natural language processing tools can find patterns in data that may not be immediately apparent to us, because of the sheer volume of data they can process. A stark example of this was the use of one such tool to find an atypical pattern of pneumonia cases in Wuhan, China, in what was the first recognition of what would become known as COVID-19.
However, this technology still has some drawbacks, namely the complexity of human language. This includes not only the language itself but also contextual nuances and even misspellings, all of which can lead to incorrect output data. Other important issues with this AI tool are the quality of the input data and the time it takes to train it. This means that for very niche subjects, it may be necessary to develop a custom-made natural language processing tool, which leads us to the final disadvantage: these tools are very costly to make.
Regulatory Compliance
Anyone who works in clinical research knows that keeping accurate and compliant records is paramount, but also extremely time-consuming. Many companies still employ a lot of people whose function is to keep track of records, their completeness and compliance. Other companies have opted for using AI-powered tools to automate the process of keeping tabs on the documents that are being filed and to analyse data regarding their regulatory compliance. These tools can quickly raise an alarm when there are incomplete documentation and/or potential compliance issues, allowing the (human) team to investigate the potential issues.
Despite not being perfect, these tools can be effectively used to reduce the time teams spend on trivial issues, freeing their time for the more important ones.
Making Companies More Human
No, this is not a mistake, but more of a paradox. Across the life sciences field, the use of AI can help professionals to save time in repetitive tasks, and processes that take too long (like analysing hundreds of research papers), feeing not only their time but also their cognitive power for things that AI is not good at: establishing connections with humans. Life sciences professionals can make a better use of their time if they can use AI findings to extrapolate their meaning, deduce the better path forward, and forge creative solutions for the difficult problems that invariably appear.
By allowing AI tools to focus on analysing giant datasets, process data and automate workflows, companies can then—through their human teams—become more personalised in the way they collaborate with their clients, healthcare professionals, patients, and other stakeholders, thus becoming more human.
References:
- Liu L, Miguel-Cruz A. Technology adoption and diffusion in healthcare at onset of COVID-19 and beyond. Healthc Manage Forum. 2022;35(3):161-167. doi:10.1177/08404704211058842
- How Canadian AI start-up BlueDot spotted Coronavirus before anyone else had a clue. Published March 10, 2020. Accessed July 30, 2024. https://diginomica.com/how-canadian-ai-start-bluedot-spotted-coronavirus-anyone-else-had-clue
- Bogoch II, Watts A, Thomas-Bachli A, Huber C, Kraemer MUG, Khan K. Pneumonia of unknown aetiology in Wuhan, China: potential for international spread via commercial air travel. J Travel Med. 2020;27(2):taaa008. doi:10.1093/jtm/taaa008
- Harrer S. Attention is not all you need: the complicated case of ethically using large language models in healthcare and medicine. eBioMedicine. 2023;90:104512. doi:10.1016/j.ebiom.2023.104512
- Patil RS, Kulkarni SB, Gaikwad VL. Artificial intelligence in pharmaceutical regulatory affairs. Drug Discovery Today. 2023;28(9):103700. doi:10.1016/j.drudis.2023.103700