Last year (2024) saw the pharmaceutical industry take significant steps towards using AI more systematically and effectively in drug research.
In 2025, access to ever-more powerful AI systems – including life sciences domain-specific Generative AI tools – will allow researchers to leverage AI to design novel drugs.
This is backed by research conducted earlier this year that showed 94% of researchers expect/believe AI will help to accelerate knowledge discovery.
However, for AI to be used effectively and safely in R&D, it will be vitally important that AI systems are built on a foundation of trusted, reliable and comprehensive data sources that are domain-specific for drug discovery and development.
General-purpose AI models are unable to meet the needs of the scientific community, which cannot run the risk of hallucinations or out-of-context predictions that are simply wrong — and lead to expensive mistakes — such as launching clinical trials with a compound for which the likelihood of success has been wrongly predicted.
To counteract this risk, retrieval augmented generation architectures (RAGs) will become the norm for AI use in sensitive industries.
By limiting the data generative AI tools can draw from to context and domain-specific sources, researchers will be able to ensure outputs generated by AI are accurate and trustworthy.