2023 is the year in which the general public learned about the unimaginable creative possibilities of Artificial Intelligence (AI). In the pharma and biotech industry, the technology is truly heralding the beginning of a new era. With Andrii Buvailo, pharmaceutical industry analyst and co-founder & editor at BiopharmaTrend.com, we take a closer look at the impact, challenges and incredible opportunities that AI presents for the sector. Andrii Buvailo will be keynote speaker at the fully booked annual event of bio.be/essenscia: ‘Shaping the future of biotech in Belgium: what’s next?’.
Mister Buvailo, how has AI already transformed the drug and therapy discovery process?
“The impact of recent AI advances goes deep. For instance, it is now becoming more realistic to model biology at scale, integrating a multitude of experimental datasets – genomics, transcriptomics, proteomics, metabolomics, epigenomics, and phenomics (shortly referred to as ‘omics’). For instance, an Israeli-based company applied AI to create sophisticated disease models for target discovery drug design, as well as informing better preclinical and clinical research.
In another use case, AI is being used by companies to design novel drug candidates for known and novel targets. Over the last couple of years, dozens of AI-designed drug candidates were advanced into clinical trials, especially in oncology, neurology, rare diseases and COVID-19.
Perhaps, the most developed use case for AI adoption is in clinical research, where it is used to streamline patient recruitment, predict treatment outcomes, integrate real-world data, and enabling more flexible and cost efficient biomarker-led trials with personalized medicine in focus.”
In what ways does AI accelerate the research and development timeline for new drugs and therapies? And does AI also play a decisive role in developing new biomanufacturing processes?
“Over the last several years, AI proved to be capable of accelerating de-novo drug discovery. For instance, it took under 30 months for a biotech company in Asia to discovery novel target relevant to fibrosis, design novel drug candidate and advance it to phase 1 clinical trial – at least twice as fast as a traditional preclinical program. Similar rapid projects were reported for other AI-driven companies. At the same time fascinating AI-enabled results are achieved in drug repurposing.
Probably the most vivid example is the success of Moderna, an inherently ‘digital-first’ company. They implemented AI-driven digital processes at every step of R&D and production, and when the time came, this helped them develop COVID-19 vaccine in an unprecedentedly short period of time – within 65 days from start to dozing the first subject.
AI is certainly playing the increasingly decisive role in biomanufacturing. Traditional pharmaceutical manufacturing of small molecules-based drugs predominantly employs legacy, retrospective analyses with limited scopes. Amidst the shift towards biologics and biomanufacturing, as data related to these complex molecules grows, the need intensifies for real-time, holistic analytics, a demand unmet by legacy systems. This all is driving a shift towards “Pharmaceutical Industry 4.0,” with not only automated manufacturing processes, but also integrated cognitive functions, enabled by AI.”
While AI offers numerous advantages, what are some of the challenges and limitations that the pharmaceutical and biotech industry should be aware of when implementing AI technologies?
“Companies should realize that while the application of predictive and generative AI models in modern drug discovery is already proving impactful, the immediate practical value varies dramatically from one use case to another. Sometimes, implementing complex AI-driven tools is critical to gain strategic or operational advantage. Other times, it is an overkill and simpler methods can be used just as effectively to deliver robust results.
One critical aspect that limits the AI adoption in many areas of life sciences is lack of quality data with proper annotation. Having the right kind of data with the right kind of infrastructure for handling and modeling such data is just as important, as having cutting-edge AI algorithms.
This brings to another important aspect to think about: the overheads of AI adoption. While AI is widely advertised as a cost-saver (which it may very well be, strategically speaking), it should be clear that having all the digital infrastructure for large-scale AI-modeling projects can be quite expensive, with lots of recurrent payments for cloud servers, computing power, maintenance, frameworks, models, etc. Using third party solutions may run into hundreds of thousands, millions and even tens of millions of dollars per year.
Also, there are issues with AI explainability, bias, regulatory aspects – especially in the European Union. All this adds to project complexity and cost.
Human resource strategy is central to the AI adoption, as AI talent is scarce and expensive, especially multidisciplinary type. Leading tech and life science corporations are headhunting for the best of the best, making smaller companies’ quest ever more difficult.
Last but not least, AI implementation requires a cultural shift on the organizational level. While everyone likes innovation, people tend to be inert when it comes to change, especially a rapid and comprehensive change. So, a proper re-skilling strategy should be taken into account, with possible focus on creating centers of excellence, running internal workshops and masterclasses, and other ways of supporting employees in their transition towards AI-driven processes and operations.”
Looking ahead, how do you envision AI’s role evolving in the pharmaceutical and biotech sector over the next decade? What transformative changes do you foresee, what challenges might need to be overcome to realize this vision and what opportunities could lay ahead?
“I think in the next decade, AI will be a cornerstone of the pharmaceutical and biotech sector – in research, biomanufacturing and the entire pharma value chain. Speaking about R&D, the drug discovery success rate is likely to increase from current almost gambling rate to more decent numbers. We will most probably see less drug failures in clinical trials, because biology models will become more accurate. As AI integrates more deeply into biotech, opportunities for real-time patient monitoring, adaptive clinical trials, and next-gen smart therapeutics will emerge, marrying technology with biology in unprecedented ways. The industry will largely shift to personalized medicine and therapies.
As Pharma 4.0 takes shape, AI will help optimize production lines, enhance pharmaceutical supply chain efficiency, and enable predictive maintenance, reducing downtimes and costs. However, these advancements will necessitate overcoming data integration challenges, ensuring interoperability across platforms, and navigating evolving global regulations. Also, a global educational shift will be needed to account for growing need for interdisciplinary talent, and business leadership able to implement deep tech projects.”
