Assessing the Role of AI in Drug Discovery
From hype to reality?
AI (Artificial Intelligence) has long been used by the life sciences industry, but it appears to have reached an inflection point in 2023. More than a sudden attraction to future-facing innovations (in an industry that has historically been slow to adopt novelties), drugmakers are facing mounting pressures to discover drugs and bring them to market faster and cheaper. AI therefore appears as a lifeline to help overcome long-standing challenges like high failure rates and lengthy development timelines. The previous year saw a stampede of startups coming into the life sciences industry with AI and ML (Machine Learning) tools that are (finally) fit for purpose, leading the AI market in healthcare and life sciences to double in size to become valued at US$20 billion in 2023 compared with US$11 billion in 2021. In their quest for efficiency, accuracy and speed, 2023 saw major drugmakers betting on AI to accelerate drug development. And this trend is here to stay.
“AI has been around for some time and is already in use in the pharmaceutical industry, but what is exciting about this moment is that AI has become democratized – meaning, you do not need a Ph.D. in programming and advanced analytics to access this technology.”
Greg Rotz, Transformation Consulting Leader for Pharmaceuticals & Life Sciences, PwC
From the drawing board to the Boardroom
From recognizing the potential to realizing it, Big Pharma and other large biotechs shifted AI from the drawing board to experimenting with it for their lead candidates. Many of the executives interviewed have identified the acquisition and implementation of innovative products (such as AI platforms) as key strategic objectives going forward. Sanofi launched its own AI app in 2023, to become what CEO Paul Hudson called, “the first pharma company powered by artificial intelligence at scale”. Roche plans to build its own AI tool, dubbed RocheGPT. Pfizer, AstraZeneca and Boehringer Ingelheim all made hundred-million-dollar investments in target and lead identification. AI-powered discovery tools have matured to the point that leading biotechs are confident leveraging them for their lead assets. CAR-T-focused Arcellx is increasingly using AI at the higher levels of drug discovery in its mission to revolutionize cell therapy. Its CBO, Aileen Fernandes, explained the potential: “We are actively considering how it can augment our efforts in drug development and bolster our pipeline. We are particularly seeing its potential in target identification, especially concerning synthetic domains and optimized targets.”
Perhaps the most harrowing example of AI’s growing importance in the life sciences industry is its transition from the drawing board to the Boardroom. Nearly every executive mentioned AI during their JP Morgan Conference speech in January 2024, and over a dozen interviews with pharma and biotech executives suggest that AI is now CEO territory, compared with a few years back. Rohit Vashisht, co-founder and CEO of WhizAI, recalled the growing involvement of C-level executives: “Previously, AI was often perceived as complex and esoteric, hindering widespread adoption across organizations. C-level executives now recognize AI's potential to drive innovation, enhance operational efficiency, and gain a competitive edge in the market.”
An AI-designed FDA-approved drug remains unrealistic in the near future, however, AI will undoubtedly make more strides in optimizing the properties of molecules, leveraging predictive data, and helping drug candidates meet target product profiles. But integrating AI in drug discovery is a complex endeavor, not least given the stringent regulations. When asked about the timeline for the approval of an AI-generated drug, Karen Lackey, CEO of small molecule drug discovery acceleration platform and DNA-encoded library (DEL) technology pioneer X-Chem, forecasted: “I do not foresee it happening until we reach the era of quantum computing. Currently, human intervention is necessary to set up and interpret data, ask pertinent questions, and gather necessary information for an Investigational New Drug (IND) application.”
“AI should expedite the process to deliver actionable insights within five years, leveraging clinical data to identify shortcomings in drug development. Once this process is optimized, the integration of quantum computing could further enhance AI capabilities, potentially leading to the discovery of AI-driven drugs.”
Karen Lackey, CEO, X-Chem
Reinventing with GenAI
There is no doubt that GenAI (Generative Artificial Intelligence), notably deep learning models and networks, is revolutionizing how pharma firms can develop drugs and bring them faster to market. In recent years, the business case for GenAI has moved from one of interrogation to one of democratization in adoption; a McKinsey report of June 2023 highlighted that “the fastest adoption of generative AI is in high tech, banking, pharma, medicine and life sciences.”
2023 was a pioneer year for GenAI across the life sciences spectrum: Biotech Adaptyv Bio leveraged GenAI to launch a protein engineering foundry to develop new medicines, and University of Central Florida researchers introduced an AI-assisted technology for drug-target affinity projections. Insilico Medicine developed a drug candidate using a tenth of the usual cost and a third of the time. To sum up the high-impact end, executives can leverage GenAI for faster drug molecular design (up to 25% reduction in production period, according to a BCG report), accelerated clinical development that could cut down writing time by a third, and enhanced quality management. In many ways, 2024 and 2025 will be about moving from hype to reality for GenAI’s applications in life sciences, and building trust amongst regulators, innovators, and pharma executives.
We are not yet at the dawn of an AI-discovered drug. Pharma companies are more likely to leverage the tool to augment the human scientific experience, rather than to replace it. When asked if AI itself could discover drugs, ChatGPT, the most popular artificial intelligence app, humbly answered: “While AI offers promising opportunities in drug discovery, it is important to note that it is not a replacement for human expertise but rather a tool that can enhance human capabilities in the complex process of discovering and developing drugs. Collaboration between AI systems and scientists remains crucial for successful drug discovery.”
Article header image by Siarhei at Adobe Stock