When It Comes to AI in Drug Development, Should We Believe the Hype?

Today’s drug discovery landscape is notoriously treacherous. Only about 10% of novel therapies ever make it from bench to bedside – and that’s at an average cost of almost $3 billion, with development timelines stretching out more than a decade. 

Understandably, the biopharmaceutical industry is heavily invested in improving these numbers, and has been looking at the potential of technology innovations like artificial intelligence (AI) and machine learning (ML) to help close the translational gap, or valley of death, that persists in drug discovery R&D. 

To explore the true transformative potential of AI in drug development, VeriSIM Life’s Chief Scientific Officer, Szczepan Baran, convened a panel of industry leaders to take a deep dive on the question: “AI in drug development, hype or reality?” 

Joining Szczepan for the hour-long panel discussion were:

  • Andrea Derix, R&D Global Program Management, Bayer 
  • Peter Henstock, Machine Learning & AI Technical Lead, Pfizer
  • Christos Varsakelis, Associate Director AI/ML, Janssen 

The AI advantage in drug discovery

AI technologies like deep learning, machine learning and natural language processing have the potential to address many of the challenges that traditionally plague drug R&D – accelerating molecule design and testing, streamlining essential processes, improving chances of clinical success, and reducing costs throughout the development pipeline. 

In trial design, for example, AI-driven tech is being used to automate trial operations – finding the most beneficial patient subpopulations, analyzing real world data like social media content to find condition-specific cohorts, and identifying target locations. Successful use cases for AI-based models are also emerging for drug repurposing efforts, where AI and ML techniques are helping identify new indications for existing licensed drugs. Other areas panelists noted as being centers of AI innovation included personalized medicine and bioprocessing. On the latter, Christos Varsakelis explained, “I think bioprocessing is one of the areas where a business case [for AI in drug development] can be made. So we have estimates right now that can show some in silico methods can drive down development timelines by 50%.”

Barriers to adoption of AI

But even as AI-driven technologies are showing their merit and global R&D investment in AI breaks records (we’ve seen a 22% increase in R&D investment in AI over the last 3 years by top-15 global pharma, with $2.1 B in revenues projected from AI-based solutions in 2022 alone), Szczepan Baran explained that decades-old R&D roadblocks and technology skepticism are still slowing the overall pace of adoption of AI technology in drug development. “Our tolerance for risk when we are engaging novel technologies is significantly lower [than for established technologies].” 

From AI’s black box perception problem, to fear of human replacement by AI, to the need for more cross-disciplinary experts in data science and biology, there are a range of additional factors acting as barriers to an industry-wide AI revolution. On the topic of the AI talent wars and the importance of upskilling and reskilling employees to meet demand, Peter Henstock remarked, “For us, we are actually running training classes for artificial intelligence and machine learning [...] We followed it with a Python course and a deep learning course, and now we’re looking at better training for the statistics group, and how we can leverage their skills in the AI space.”

Better AI benchmarking

Effective evaluation of AI applications within drug discovery has been a struggle for the biopharma industry, as traditional endpoints like time and cost are not always the right variables for measuring AI’s total impact. As Andrea Derix explained, “I think everybody is looking at time and resource spending, but not as much at how the quality of decisions, or quality of development, can be enhanced by AI.” 

To move past the hype to do a better job of benchmarking AI’s real successes, panelists agreed there should be more focus on tracking and promoting effective AI use cases, which would help build trust and gain buy-in from decision makers and stakeholders within organizations – ultimately creating more opportunity for an AI-driven paradigm shift in drug discovery.

VeriSIM Life’s BIOiSIM platform and unique Translational Index™️ technology are helping close the translational gap in drug R&D. Watch our full panel discussion: AI in Drug Development: Hype or Reality or contact us to learn more about AI-powered drug development.

Learn more about VeriSIM Life’s BIOiSIM platform and unique Translational Index™️ technology.