How AI is accelerating the development of new treatments for specific patient populations

Make sure to listen to the full episode on any of the following platforms:

• SAS
• Spotify
• Apple Podcasts

Artificial Intelligence (AI) will only grow in importance for the health and life science industry. We already see it as really valuable for helping us develop new treatments for people – and faster. Before I delve into the possibilities and challenges that AI brings, let me take you back to when I first became interested in AI.

Watson intrigued me
I remember seeing one specific episode of the Jeopardy! game show on TV at some point during 2011. IBM’s Watson computer beat America’s best players. That intrigued me. Just a month later, I got chatting to John Kelly – IBM’s Head of Research and father of Watson – at a conference. We talked excitedly about the possibilities of AI.

By the time a few months had passed, we had already built a joint AI project. Watson used cognitive computing – specifically patient similarity analytics and real-world data – to predict potential outcomes and identify personalized care for people living with epilepsy.

We quickly saw incredible enthusiasm for AI grow across UCB. As a result, we applied for a couple of patents, primarily as recognition of our employees’ great work.

We wanted to bring every UCB employee with us on our AI journey. So, in 2016, we created a technology practice, a practice that later became our ‘DataMinds and DigiMinds program’. The program helped increase digital understanding and capabilities at UCB, specifically around data literacy and digital business transformation.

I put the fantastic number of AI initiatives we’re seeing at UCB today down to that program, initiatives that span the whole value chain.

Amazing AI initiatives
You only have to look at our Moonshot initiative, a joint initiative we worked on with Microsoft last summer. Combining our AI algorithms with Microsoft’s massive compute power allowed us to research potential treatments for COVID at an accelerated speed. Thanks to AI, we completed research that would typically take six months in just three days.

More generally, I’m convinced that AI can play a significant role in developing the COVID vaccines and potential treatments we see today. But that’s not specific to COVID; I’m seeing the whole pharma industry embracing AI.

Looking to the future, I believe there will be massive – and sometimes scary – advances with AI on many fronts. Take GPT-3, a language model that uses deep learning to produce human-like text. GPT-3 already uses 175 billion machine learning parameters, ten times as much as the technology we mainly see today. Putting such capacity and capability into a robot is just huge.

Proceeding with caution
I also see AI becoming more embedded in technology products in such a way that it won’t always be obvious what’s AI and what’s not. So there needs to be a set of rules around that, and the European Commission is working on those.

Other significant issues around AI need to be addressed, including ethical issues, data privacy issues and bias in both the algorithm and the data. Every dataset, every algorithm inherently has a bias, which means every AI initiative is based on something biased. You need to be aware of it to overcome it, and we do that by including bias as one of the elements we look at in our ethical reviews.