Sanofi, a leading French pharmaceutical company, has introduced an innovative artificial intelligence (AI) application called PLAI that has the potential to revolutionize the drug discovery process. With access to over 1 billion data points, PLAI provides concise and timely information to Sanofi’s employees, ranging from inventory warnings to meeting suggestions and clinical trial site recommendations.
The CEO of Sanofi, Paul Hudson, enthusiastically showcases PLAI, comparing its functionality to Netflix recommendations. Hudson jokingly boasts that the app broke even within four hours of its launch and emphasizes its affordability compared to the expensive fees charged by consulting firms for data management projects. Roughly 10% of Sanofi’s 80,000 staff members utilize PLAI on a daily basis.
While AI has been utilized by biotech firms for years, it is now garnering increased interest from major pharmaceutical companies. Emma Walmsley, the CEO of GSK, believes that AI can greatly enhance the productivity of research and development, which has been a significant challenge within the industry. Moderna, another prominent pharmaceutical company, has also expressed its strong focus on AI adoption. Sanofi is fully committed to leveraging AI technology, and Morgan Stanley predicts that the pharmaceutical industry could be spending $50 billion per year on AI within a decade to accelerate the drug development process.
The main excitement surrounding AI in drugmaking lies in its potential to significantly improve the hit-and-miss nature of drug discovery. Currently, it takes approximately a decade to bring a drug to market, costing billions of dollars with a success rate of only 10%. Even a small advancement in efficiency and speed could be invaluable. However, scientists have faced challenges in analyzing the vast amount of biological data using traditional statistical methods. Machine learning has emerged as a solution, allowing researchers to analyze extensive datasets, including clinical patient data, genome sequences, and medical imagery. DeepMind, Google’s AI lab, recently made a significant breakthrough by using its AlphaFold system to predict the structure of nearly all proteins, which could assist in identifying molecules with therapeutic potential in the future.
Although only a limited number of drugs in development have thus far incorporated AI, this list is expected to expand rapidly, particularly for simpler molecules that are relatively easy to predict. In the realm of these straightforward chemistries, the future of medicine appears increasingly computational.
AstraZeneca, a leading British pharmaceutical company, utilizes AI technology in 70% of its small molecule projects. Through a technique known as reinforcement learning, AstraZeneca’s AI continually refines its molecular suggestions and simulates how modified molecules might react. Another biotech startup, E-therapeutics, leverages AI algorithms to design RNA molecules based on the sequences of all the genes in specific organs, allowing for predictions of the molecules’ activity and their potential in inhibiting disease-causing genes.
Another promising application of AI is the use of knowledge graphs, which store data about genes, proteins, diseases, drugs, and the interconnected biological pathways. These knowledge graphs can aid in identifying new targets for drug development. Additionally, generative AI is being explored to suggest entirely new chemical and biological structures for testing, similar to how ChatGPT can generate unique pieces of content. Beyond drug discovery, AI applications such as PLAI could address significant challenges of efficiency in the heavily regulated and labor-intensive pharmaceutical sector.
Despite the promising potential, some pharmaceutical executives express concerns about generative AI’s tendency to generate unreliable information and potentially lead researchers astray. In fact, Hudson reveals that many CEOs he has spoken to harbor fears about the existential threats posed by AI. However, he envisions the next industrial revolution rather than a robot uprising.
To stay informed about the latest business and technology news, subscribe to the Bottom Line, our exclusive weekly newsletter.