Under The Scope: Harnessing the Power of GenAI in Life Sciences

Welcome to the 6th issue of #UndertheScope, Gilbert Meher | Science's newsletter designed to bring communities together in Science.

In this edition, Emily Legood sat down with Tyrone Richardson, the Chief Commercial Officer of Ryght, who is at the forefront of integrating generative AI into life sciences sector, talking about the potential of GenAI in the #biotech, #pharma, and #CRO sectors.

Based on current trends, the global AI in life science analytics market size is expected to reach around $5.10 billion by 2032. This will be a significant increase from the $2 billion recorded in 2023, with a Compound Annual Growth Rate (CAGR) of 10.98% during the forecast period. [Source: precedenceresearch.com]

Data Challenges in Life Sciences

Electronic health records, biobanks, genomic studies, patient registries, and imaging & digital pathology: data in the life sciences and healthcare industries is increasing in volume and diversity as a result of the digitalisation efforts in the past decade. As technology evolves, the market is becoming rich in data, but much of it is unstructured, making it very difficult to analyse.

Generative artificial intelligence (generative AI, GenAI) is artificial intelligence capable of generating text, images, videos, or other data using generative models, often in response to a user’s prompt or request. - IBM

GenAI uses machine learning models, especially deep learning techniques, to analyse and learn from large corpus of data. These models can then create new content or make predictions based on the patterns they've learned.

In the field of life sciences, generative AI handles large volumes of unstructured data, like medical records, genomic sequences, and clinical trial data, to extract meaningful insights and generate relevant outputs.

"GenAI is very good at looking at text and large amounts of unstructured data sets to unlock the data capabilities of these organisations. This will help accelerate drug discovery, optimise clinical trial designs, and speed up the process towards customised treatment plans for individual patients." - Tyrone Richardson

Why Generative AI?

Tyrone listed the 3 main benefits of using generative AI:

1. Search and Retrieval

Generative AI is particularly good at searching large datasets and retrieving relevant information quickly, making it easier for users to sift through vast amounts of data and find what they need, fast.

2. Content Generation

It is highly effective in content generation - it helps create the framework of essential documents and tackle the challenge of the "blank page problem" that many of us encounter. This is particularly useful for small to medium-sized biotech companies where staff often wear multiple hats.

3. Workflows

Broad tasks like research or documentation generation can be time-consuming and unstructured. AI can stitch together components of extensive tasks in end-to-end workflows, which accelerates and uncovers insights that would otherwise be missed.

"For example, imagine a professional in the life sciences field being able to search databases like PubMed and sift through extensive clinical published data with the help of GenAI. They can then analyse or cross-compare their own data sets to gain further insights and generate content in a streamlined workflow", Tyrone explained.

"The concept is that AI can handle 80% of the time-consuming foundational work, freeing up experts to focus on the critical final 20%." - Tyrone Richardson

Challenges of Integrating AI into the Sector

1. Data Security and Privacy:

The security of sensitive data, including an organisation's intellectual property and healthcare data, is crucial. Organisations need to implement strong security measures to prevent data breaches and protect confidential information. This includes using encryption, secure data storage, and strict access controls.

"Life sciences and Healthcare organisations should be able to work with their own data in a secure and private way. For pharma and biotech, intellectual property is the value of the company, and should be protected. " - Tyrone Richardson

2. AI Hallucinations:

AI hallucinations happen when a generative AI model produces outputs that are either nonsensical or well-written yet factually incorrect. This can occur due to biases in the training data, overfitting, or the model's inherent limitations. To reduce hallucinations, it's important to use high-quality, diverse training data and have convenient validation processes in place.

"At Ryght, we've gone to great lengths to minimise hallucinations by allowing users of our platform to check the sources of generated responses. The users can cross reference and approve the accuracy on a factual level." - Tyrone Richardson

3. The Risk of Technical Debt:

Technical debt refers to future costs incurred when quick, easy solutions are chosen over more complicated approaches that take longer to implement. In AI, this can happen if the models aren't continuously updated or if the system isn't scalable. This often results in higher future costs and inefficiencies.

"The risk of technical debt is high, especially when organisations develop their own AI platforms, which may not be enterprise-grade. Meanwhile, AI models evolves quickly, and without continuous updates, companies can fall into technical debt." - Tyrone Richardson

What does Ryght do?

"We saw the generative AI revolution, and have the experience to know the technology needs to be secure and specific to life sciences to have success. We are democratising the technology to allow organisations to access it securely and quickly. We're aiming to transform how life science professionals create documents, surface the latest research, grow revenue, and get drugs to market faster." - Tyrone Richardson

Ryght is on a mission to make generative AI accessible, secure, and tailored for life sciences and healthcare applications. They offer a secure enterprise-grade generative AI platform with a growing arsenal of copilots that help sites, sponsors, and CROs run more efficiently.

To address the challenges of integrating AI into the life sciences sector, Ryght ensures enterprise-grade security to protect sensitive data. This means individual data remains private and isn't used to train other language models. Their platform uses a loosely coupled architecture to integrate various AI components and models, making it highly scalable and always updated with the latest advancements in AI technology.

Vision for the Future

Tyrone sees AI continuing to evolve rapidly, becoming even more integral in processing vast amounts of complex data.

"I think we're going to continue to see different AI and ML models, whether they're generative AI or algorithmic prediction models, continue to move forward and help us reason with information that would otherwise be too difficult or time-consuming to deal with." - Tyrone Richardson

Looking ahead, Tyrone envisions a future where AI reduces the time and cost of bringing new medications to market, personalises treatment plans, and empowers research and development.

"We hope to see a future where medications are matched to patients more efficiently, improving therapeutic outcomes and reducing costs," Tyrone shared. The goal is to make clinical trials more efficient and ensure innovative treatments reach patients faster.