AI for living beings: fiction and reality
-
Myth No. 1: "AI will never be able to diagnose a disease reliably."
In certain areas, such as the analysis of medical images or genetic data, AI can already help to detect diseases. But it does not diagnose on its own. It assists doctors, who interpret the results and make the final decision. Its reliability depends above all on the quality of the data and the methods used to train it, an area in which bioinformatics plays a key role.
-
Myth #2: "The more data an AI model is trained on, the more reliable it is."
AI models trained on very large amounts of data attract a lot of attention, but they are not always the most accurate. In the life sciences, more specialised models can sometimes analyse genetic, medical or environmental data more effectively. Bioinformaticians rigorously compare and test models – a process known as benchmarking – to identify the most reliable methods.
-
Myth n°3: "AI is necessarily neutral".
An AI learns from the data it is given. If this data is incomplete or erroneous, the results may be biased. For example, a medical tool trained on certain populations may be less reliable for others. In bioinformatics, experts check, organise and document the data used to avoid bias as much as possible and make the analyses more reliable.
-
Common misconception No. 4: ‘AI is a recent revolution in the life sciences.’
Scientists have been using artificial intelligence and machine learning methods for decades to analyse genomes, proteins and medical data. Recent advances – notably generative AI tools such as ChatGPT – have propelled AI into the spotlight, opening up new possibilities and accelerating research. However, its reliable application in the life sciences relies on the body of accumulated scientific knowledge, as well as on high-quality biological databases, developed and validated through bioinformatics.
-
Common misconception no. 5: ‘With AI, scientists no longer need to carry out experiments.’
AI can be used to detect early signs of disease, predict the course of an epidemic or guide biodiversity conservation efforts. But it does not work alone: it relies on vast quantities of data produced by scientists – in laboratories, in hospitals or in the field. These observations and experiments remain essential for evaluating and improving AI models, as well as for generating new knowledge.
-
Bioinformatics: the foundation and driving force behind trustworthy AI in the life sciences
In the life sciences, bioinformatics is both the foundation and the driving force behind artificial intelligence. It organises and enriches biological data, develops and uses AI models to analyse it, and rigorously compares methods to verify their reliability. This expertise enables biological data to be transformed into scientific discoveries that benefit society, from healthcare to the environment.