PING 2024 round-up - 'AI in pharma - Threat or opportunity?'
Over a month has passed since the 2024 PING conference, providing invaluable insights into the uptake of AI in Pharma, its transformational promise, and the accompanying risks and pitfalls. Below is a summary of our key takeaways.
This year's PING (Pharmaceutical Industry Network Group) conference explored the rapidly growing role of artificial intelligence in the pharma industry. Delegates heard from various speakers operating at different points in the pharma lifecycle, from R&D to supply chain management, on how AI is transforming their work and creating opportunities.
Below is a summary of the key points from the talks and identifies particular areas of interest for those working within the pharma industry.
Session 1/7 - Overview of AI in pharma - threat or opportunity?
Dr Andrée Bates of Eularis, a leading AI consultancy focused on the pharma sector, kicked off the talks with a broad summary of what’s happening in the AI world in relation to pharma.
AI has shown promise in understanding and managing regulatory dossiers and guidance documents, automating medical legal reviews and strategic literature reviews, faster reimbursement, preparing for launch, finding rare disease patients, and much more.
Session 2/7 - Is AI transforming drug discovery and product development for the better?
For the second session, a panel of experts from across the pharma industry discussed how transformational AI could be in drug discovery and development. Darren Spevick of Russell Strategy Partners chaired a panel of experts: Nicola Richmond (BenevolentAI), Gareth Langley (Charles River Laboratories), Nick Street (Healx) and Jordan Lane (Ignota Labs).
In terms of approaching the use of AI, interdisciplinary collaboration was identified as being vital to facilitate the integration of AI tools with human expertise, fortifying trust in AI-driven decisions. To build a robust foundation for AI in pharma, data quality and transparency of confidence ratings in a dataset is key.
AI is not the silver bullet to drug discovery. It can move the dial for successfully converting candidate drugs, reduce costs and play a role in repurposing and refining distressed assets.
Challenges to AI adoption
To bridge the significant gap in AI adoption, increasing knowledge and understanding of AI's opportunities and threats, enhancing data comprehension, and democratising AI technology will be vital. Democratising the technology is a particular challenge, as big pharma is drawn towards developing its capabilities internally.
The panellists acknowledged the inherent bias in the drug discovery process. Those interested in harnessing AI should be particularly aware of the potential for bias and adopt strategies to address this, for example, by ensuring that training data is fair and balanced, and accounting for bias when assessing outputs.
Looking forward
The panellists envisaged a world where AI and human intelligence harmonise to streamline R&D, lower operational costs, and significantly improve clinical decision-making accuracy.
To facilitate this, it was discussed that current R&D norms need to be challenged, so that data needed to avoid bias is collected, and siloing of data be avoided through promotion of startups, partnerships, and innovative risk-sharing models.
However, expectations of AI's speeding up of quality assurance and efficiency gains are in the wrong place. Biological processes take time and the feedback loops are longer in pharma than in other sectors to ensure safety. Pharma companies should bear this in mind and manage expectations when taking steps to adopt AI in their work.
Session 3/7 - AI to help with the holy grail of personalised medicine
In the third session of the day, Dr Sola Adeleke, founder and CEO of Curenetics, spoke in conversation with VWV's Sonya May to discuss how AI can assist with the development and acceleration of personalised medicine.
Dr Adeleke comes from an oncology background, giving him an insight into the issues that arise with common cancer treatments like radiotherapy and chemotherapy, and the difficulties in predicting how a patient will respond to a particular drug.
Growing in-house talent to develop AI tools which analyse and overlay biomarkers in genomic data, data from patients' scans and other clinical data to predict how patients will respond to treatment has proved vital.
Dr Adeleke shared that Curenetics hopes to get a product to market within the next 6-12 months, marking an important step for personalised medicine and cancer treatment.
Session 4/7 - Using AI: Is It Legal?
Our fourth session of this year’s PING conference scratched the surface on the lawfulness of using AI, and was presented by Harry Jennings of VWV's Pharmaceuticals and Life Sciences team.
One of the most critical questions that has emerged with use of AI is how to navigate uncertainties relating to intellectual property rights. It’s crucial to understand the ownership of AI-generated content and the data underpinning these systems; there are many unanswered questions. With AI patents on the rise and several high-profile copyright disputes around data scraping for AI training, it’s evident that we're at the cusp of significant legal evolution.
The New York Times v Open AI and the Getty Images v Stability AI cases have highlighted the urgent need for clarity for AI developers and users. There is already some regulation in place, and more on the way. However, as the UK government at the time of the conference was adopting a 'wait-and-see' approach, we find ourselves in a realm of uncertainty and anticipation. It is yet to be determined how our laws will adapt to the fast-paced technological advances relating to AI. For now, some of these points will be considered and addressed in the courts … if they get that far.
Harry put forward the view that the owners of materials covered by copyright, trade marks and database rights will need to be compensated for the use of their IP to train AI systems. And it's possible that AI businesses will be required to provide a transparency statement on how their AI was trained, and on what data, to help identify risks in the AI, such as the risk of biased outputs.
In other areas, AI raises important questions about the application of law to this new(ish) field. Two clear examples are: liability arising from products or services using or relying on defective AI, and data protection compliance.
To discuss these topics in further detail, Harry will be running a series of talks and meetings, aiming to unravel the complexities of AI law. The series is called “Make Me Intelligent”. Please let us know if you would like to sign up to the Make Me Intelligent list.
Session 5/7 - Can AI be used to solve healthcare disparities?
In the fifth session of the day, Dr Rav Seeruthun, co-founder of health-equity.ai, spoke about how they are using AI to combat healthcare disparities. health-equity.ai is a medical technology company which is developing in-house AI products to lead their work in this area.
Dr Seeruthun highlighted the shocking healthcare disparities that exist within the UK today. Leveraging available UK datasets through the use of AI, health inequalities can be identified and tackled. An example given was using AI and machine learning to identify a range of factors that could predict poorly controlled diabetes, such as crime score and adult skills score. These tools are then used to adopt different clustering techniques to group patients.
As you may already know from our recent press release, health-equity.ai was the recipient of this year's PING Innovation Award to recognise their inspiring commitment to health equity. We are looking forward to seeing their future work in this important area.
Session 6/7 - Intelligent management of your pharma supply chain
Philip Ashton, co-founder and CEO of 7bridges gave a presentation on how AI can be used to make pharma supply chains more effective, flexible and resilient.
Accessing data related to a pharma supply chain can be difficult as documents may be paper-based or in different languages. Recent advances, such as LLMs, now enable us to capture and process data with greater efficiency, irrespective of the source document's form or language.
Small changes to a pharma supply chain can have significant knock-on effects. New technology can assist with modelling and predicting responses to these changes.
AI tools can help predict where shortages or increased demand for a drug may arise, allowing pharma companies to plan appropriately and ensure that drug supplies are positioned where they need to be. As demand for certain drugs can quickly change and shortages are a source of media and public attention, many operating within distribution and supply chain management will be interested in exploring AI capabilities in this area.
Pharma companies can create 'digital twins' (essentially a digital representation of an object or process) of supply chains, enabling them to understand in advance how a pharma supply chain will respond to issues and disruptions.
Adoption of AI in supply chain management can lead to significant cost reductions, allowing pharma companies to reduce prices and increase accessibility to their products.
Given the impact of pharma supply chains on a company's P&L and customer outcomes, implementing AI solutions in this area will certainly be of interest to pharma companies, particularly as projects can pay for themselves in a matter of months, depending on their scale and scope.
Session 7/7 - Intelligent use of generative AI for competitive advantage
In our final talk of the day, James Turnbull, founder of Camino Communications, spoke to PING delegates about how Camino is harnessing AI to help pharma companies secure a competitive advantage.
To demonstrate Camino's approach, James gave an example of a project they have recently undertaken for a mid-size pharma company. This company was concerned that a competitor was moving into its space and asked Camino to review all publicly available data to see what steps the competitor had taken.
Camino used AI to review data sources such as medical journals to summarise the competitor's recent areas of focus. This was such a success that it has asked Camino to scale up its work to target all of its competitors.
This is an interesting use case that is applicable across the pharma sector and more generally in respect of competitive intelligence. It highlights the versatility and flexible of AI, and demonstrates that it is important for the industry to take an open-minded approach when identifying and developing use cases.