The pharmaceutical industry was worth $1.6 trillion in 2024, employing around 8 million people directly and up to 75 million people indirectly. Various flavours of artificial intelligence show considerable promise in these industries. AI can be utilised in various applications, including research and drug discovery, clinical trials, operations such as manufacturing, and for other purposes, such as searching medical literature and ensuring regulatory compliance.
A striking example of AI’s use in drug discovery is the company DeepMind, since acquired by Google, which has been a pioneer in “deep learning” AI techniques. It demonstrated the promise of the technology by teaching it the basic rules of chess, and then left it to play millions of games against itself, gradually improving its standard of play. The result, AlphaZero, then played a match in 2017 against the top computer chess program Stockfish. Over a hundred games, AlphaZero won 72 games and lost none. It later extended this to the more computationally challenging game of Go, where it beat the human world champion Lee Sedol in 2016. DeepMind extended this approach to the problem of protein folding, winning its founders the 2024 Nobel Prize in Chemistry. A protein’s final 3D structure determines its function, and accurate prediction of how a protein will fold allows drugs to be targeted at misfolded proteins. These are the causes of Alzheimer’s disease, cystic fibrosis and Huntington’s disease, amongst others. However, predicting how a protein will fold turns out to be computationally very hard. DeepMind’s Alphafold technology is an artificial neural network trained on a large database of molecular structures called the Protein Data Bank. The software takes the amino acid sequence of a new protein and predicts the 3D protein structure, including a confidence score of its prediction. AlphaFold was much more effective at such predictions than alternative approaches.
These theoretical developments are leading to new drugs. The first drug targeting an AI-discovered novel protein for idiopathic pulmonary fibrosis, called Rentosertib, has completed positive early clinical trials. New potential antibiotics are also appearing. Given the problems of antibiotic resistance, these are particularly encouraging, and one of them targets the very troublesome MRSA bacteria. Even though these are early days, the approach shows promise. There is a growing list of such potential new drugs.
There is a lengthy process between identifying a promising drug and it reaching the market. There is a multi-stage process of clinical trials, initially on a small sample of volunteers and then on larger groups of patients if the drug is shown to be safe and effective in early-stage trials. This is a time-consuming process, taking several years, over eight years on average in 2024. AI can help with patient recruitment by analysing their medical history and geographic location, and monitoring patients during trials and speeding up workflows such as generating analysis reports. Even a small speeding up of the clinical trials process can have a major effect on the profitability of a drug, since more sales can be made under patent protection. With the median R&D cost of a single drug being over $700 million and a clinical trial success rate of just 8%, it can be seen that even small efficiency improvements can have a significant impact.
Beyond the clinical trials process, pharmaceutical companies have to produce and distribute drugs, market treatments and carry out basic research. Various types of AI can be potentially used throughout these processes, such as helping in medical documentation searches like patent searches and scrutinising medical publications.
AI can play a role in many aspects of the pharmaceutical industry, from drug discovery to streamlining clinical trials to improvements in operations and medical research. The growing list of new drug discoveries that are now in clinical trials, in particular, is a sign that AI is having a meaningful impact on the industry. Hopefully, many of these new treatments will turn out to be safe and effective and have a real impact on people’s lives.
