Will the intersection of blockchain and AI inspire the Pharma R&D Renaissance?
Posted 30th September 2019 by Joshua Sewell
Global Engage spoke to Scott Kahn ahead of the 3rdGlobal Pharma R&D Informatics and AI Congress. Scott is the CIO of Luna PBC, the entity that manages LunaDNA: a community-owned data-sharing resource. Their conversation was recorded as a podcast, which can be found here.
This is the second half of that conversation, where Scott discusses the challenges of distributive ledger technologies such as Blockchain, and their intersection with AI for Pharma R&D.
What are the challenges that remain with blockchain?
One of its great strengths is its great weakness: the fact that the ledger is immutable.
Therefore, you cannot put personal, identifying information on the ledger because it could never be removed. One of the rights that you have at least under GDPR is you have the right to be forgotten. But how would I forget a piece of information that is immutably stored on a ledger?
A solution for this is to think through architectures that use blockchain but that also leverage other technologies. So, you hear the terms ‘on-chain’ and ‘off-chain’. Users that are ‘on-chain’ cannot share personal information that might need to be deleted. This is a strength because the data is mutable and it’s a weakness because if you want to be forgotten data needs to be deleted.
The solution is to combine these on-and-off chain approaches.
How does this intersect at this present stage where there is such an emerging interest in AI?
What AI requires are large, very diverse datasets. I would argue to be able to gather those large, diverse datasets we need to have an environment where people are comfortable running their data for it to be shared. I think what transparency provides is a mechanism whereby we can verify that whatever trust you’ve given is being held to. “Trust, but verify” is a great expression for this.
Through its described transparency ledger, blockchain provides a mechanism for individuals to constantly monitor whether or not what they had intended for the use of the data is being held to.
I’d say the way it contributes to AI is that these technologies lead to a system that can gather increasingly diverse data, and that data ultimately benefits what people want to do with all sorts of AI and machine learning.
It’s a bit of an indirect answer but that’s the way that we’re looking at it.
What are you most excited about for the future in terms of blockchain and specifically blockchain technology within R&D and drug discovery?
The thing that excites me most is the seismic shift in the way data is used, researched and controlled. We’ve grown up in a world where institutions, hospitals, academic groups, etc., control others’ data and then do whatever they wish with that data.
What blockchain provides is the mechanism where these resources are not controlled by institutions but by people. When we make that shift, some really great things happen.
We would no longer have to worry about whether institutions could transfer data and all the regulatory hurdles in place with institutions sharing data on behalf of individuals. If an individual in Europe wanted to bring their data to a place in North America, they would be completely free to do so. If you let individuals make those decisions, it becomes much more straightforward.
To summarise, the biggest thing to look for in the future is the shift from data being controlled by institutions to data being controlled by individuals: having a world in which people are nearer the centre. There are all sorts of challenges that go with that, but directionally, there are a lot of benefits as well.
The 3rd Global Pharma R&D Informatics and AI Congress will explore implementation strategies and advances in the convergence of AI, informatics and data management for early drug discovery. There’s still time to register online today.
Leave a Reply