AI and trusted data: The key to unlocking faster, more cost-effective drug discovery

The artificial intelligence (AI) landscape is evolving at a rapid pace. Generative artificial intelligence (GenAI), in particular, has taken the world by storm since the introduction of ChatGPT at the end of 2022, which brought the technology into the mainstream. It is hard to find anyone who has internet access who has not given it a try, and many companies are incorporating GenAI into their work practices and offerings.

The fast growth of these technologies is unprecedented. A 2023 McKinsey analysis suggests that, out of 16 business functions identified, there are four primary business functions that could represent 75% of the total value from generative AI use cases for business. One of these areas is research and development (R&D) – where GenAI could “deliver productivity with a value ranging from 10 to 15% of overall R&D costs.”¹

Given that pharmaceutical companies spend about 20% of their revenue on R&D, the McKinsey Global Institute estimates that GenAI alone could contribute between $60bn to $110bn a year to the industry, according to a 2024 article.² Additionally, according to recent forecasts, the global market for AI in drug discovery is expected to grow from around $1.5bn currently to approximately $13bn by 2032.³

These explosive growth predictions are fuelled by the promise of AI to transform every stage of the drug development pipeline, from target identification through to clinical trials. Pharmaceutical companies are increasingly interested in AI’s vast potential and where it can be applied next.

Transforming drug discovery and development

A particularly good example of where interest in AI is growing is the hit-to-lead (H2L) workflow. Taking a small molecule hit to a potential lead compound has long been a bottleneck in drug R&D. Designing a new molecule that can effectively interact with a target, synthesising and then testing it, consumes vast resources and time.

Technological and computational advancements have already resulted in major developments in small molecule discovery. Now, AI is emerging as a powerful tool to streamline and optimise various stages of the design-make-test-analyse cycle.

For instance, AI-powered virtual screening can rapidly evaluate millions of potential compounds, predicting their interactions with target proteins and identifying the most promising candidates for synthesis and testing. This approach can dramatically reduce the number of repeated tests required by replacing lengthy lab work with computational predictive models. Additionally, AI's ability to sift through billions of data points quickly holds immense potential for identifying novel drug targets and uncovering new therapeutic opportunities.

The benefits of AI when applied to this process have already been acknowledged in recent studies, with research highlighting how AI-based approaches can lead to more than twice the improvement over baseline on the key metric of "efficacy observed".⁴ Additionally, AI models enable over a hundred times more in silico experiments than conventional screening techniques, which aids experts to accelerate the pace of discovery. This also means AI may allow companies to do less in vivo testing and result in higher success rates.

Beyond the H2L stage, AI is also predicted to revolutionise various aspects of drug development, including preclinical testing and clinical trials. AI could enhance preclinical testing by predicting the outcomes of experiments and identifying potential safety concerns early in the process, before moving to testing on humans. Moving into clinical trials, AI can help improve patient recruitment, trial design, and data analysis of the trials.

Responsible, transparent application of AI will be crucial

There is growing recognition amongst the research community that AI will fundamentally change the pace of scientific discovery. A recent global survey of clinicians and researchers conducted by Elsevier, for example, found that the majority of those involved in research (94%) believe AI will accelerate knowledge discovery. Additionally, the vast majority of respondents said it will rapidly increase the volume of scholarly and medical research (92%) and provide cost savings to institutions and businesses (92%).

The responsible, ethical, and transparent application of AI in drug discovery and development will be a critical component of its ultimate success. It is important that pharmaceutical companies ensure their AI models are built on high-quality, up-to-date, comprehensive training data, and are rigorously validated by experts. Keeping human experts in the loop during this process, i.e., ensuring AI-generated insights are reviewed and refined by experienced researchers, is essential to mitigate risks and maximise the accuracy and safety of drug candidates.

Data silos, lack of standardisation, and the lack of skilled personnel with expertise in both AI and pharmaceutical sciences could hinder progress.

While AI has the potential to enhance and advance drug discovery, overcoming these hurdles will require collaboration between industry, academia, and regulatory bodies to establish best practices, share data, and cultivate a skilled workforce.

Looking to the future

At almost every stage of history, technology has been a disruptor. This has ultimately benefitted society through eliminating routine tasks and bringing new value and opportunity for people freed up to focus on higher-value work. Despite the challenges highlighted, the potential benefits of AI in drug discovery are significant and exciting. As the biopharma industry continues to grapple with high costs, lengthy development timelines, and the ever-present need for innovative therapies, AI offers a path forward. By harnessing the power of AI, pharmaceutical companies can bring life-saving treatments to patients faster and more cost-effectively than ever before.