The Nordic.AI Health Summit 2018 gathered researchers, industry leaders, developers, and practitioners to discuss the most prominent technology of our time in the context of modern healthcare: Artificial Intelligence (AI).
The keynote speakers included several brilliant minds from various parts of the AI-healthcare ecosystem such as Google Deepmind, NVIDIA, National Health Service (NHS), and as well as insights from political standpoints of the United Nations (UN), Human Rights Watch and more.
“Are we ready for a technological revolution?” – a thought that may appear when faced with the impact and endless possibilities of AI . Although several concerns are interlinked with AI being the new paradigm-shifting technology that alters the status quo, the Nordic.AI Health Summit empowered a different state of mind and that was assurance. Assurance that AI is not merely a hype or an attempt to replace humans but a different and new height in the way we operate and think.
“It will redefine healthcare, it won’t replace it” – Nicola Rieke, NVIDIA
The Nordic.AI Health Summit 2018 covered topics such as AI adoption in healthcare, the clinical impact of AI, and the current constraints of AI. The summit showcased how several innovators are healthcare-centric in involving technology and people.
“Everything is mobile except healthcare” – Andreas Mattson, KRY
The utilization of AI was shown in various fields, from highly scientific fields to pragmatic solutions that directly benefits the caregivers. Blueprint Genetics are working on “sequencing our genome” and uses Machine Learning probability scores to provide decision support while LeapBeyond uses AI for drug discovery.
Other approaches have been on more emergent matters such as Corti’s real-time decision support in emergency health by combining AI and decision-making or NHS digital triage algorithms when receiving emergency calls. The practical solutions provided by KRY enables remote- and online doctor visits with your phone and companies such as Ably and Nox are able to extract more data from patients through their “bed patterns” and “sleeping patterns”.
1. Grasping the capability of AI requires collaboration throughout the whole ecosystem
There is still some misconception of what AI is currently capable of and how it benefits and deals with societal problems. Dispelling the uncertainty from a technology perspective might lead to a better perceived benefit, engagement and more precise clinical impact. The acceleration of technology adoption in healthcare may increase by becoming more user- and clinical-centric. Having a larger emphasis on engaging the users is required when developing solutions.
“What’s noise to one can be a signal for another” – Dr. Egge van der Poel, CERN
As Dr. Egge van der Poel emphasizes, inequality as a result of Big Data can create opportunities in working together and knowledge sharing. Nenad Tomasev from Google Deepmind accentuates how we still have ways to go before we can close the gap between AI and Healthcare, for which we might need more interdisciplinary work and a holistic approach. Healthcare is immensely complex and a way to create more coherency in our data generation is collaborating and sharing more freely.
2. Enable data sharing, security and trust
We still face an immense issue of managing, classifying and not having access to data. Enabling better data access, knowledge dissemination thus “building bridges” allows for the generation of more clinical data which will increase our chances of improving results with AI. However, with policies and nation/region-based constraints on data security, the hurdle only increases unless a different solution for data sharing is proposed.
Trust is not given freely and this refers to patients, caregivers, policies and data itself in terms of interpretability or quality. As discussed during the panel, technology tends to be overestimated on the short-run and underestimated on the long-run. Once we are able to meet the basic needs in healthcare will AI start to shine.
3. Demand calls for a new hybrid of expertise
A culmination of the aforementioned challenges, there is a need for an entirely different approach in healthcare that combines the best of both worlds. There is still struggle when attempting to collaborate and there is even still struggle when trying to define and understand when to use the AI terminology.
“When you’re fundraising, it’s AI, when you’re hiring, it’s ML, when you’re implementing, it’s linear regression, when you’re debugging, it’s printf()” – Baron Schwartz, Vivid Cortex
However, as the technology advances and as the tasks change, so will we and adapt accordingly. Accelerating the AI adoption requires support, insight and collaboration. This is something NVIDIA have set their sights on. Their Inception program strives to propel AI startups that are advancing modern healthcare with new tools and expertise. Connecting incubators with healthcare could be the first step in nurturing the new talents with hybrid capability in the AI-healthcare endeavor.
While a steady adoption of AI is not currently in place, the stage is set. The more we interact and collaborate across the industry and the more likely it will affect the adoption of AI in healthcare. AI is not an “alien thing”, it is not even an “IT-thing” accordingly to Dr. Egge van der Poel - it is an “idea-thing” and the Nordic.AI Health Summit encourages everyone, including startups and brilliant minds from both industries, to come together so that AI may empower the caregivers.
// Johnny Le, Business Engineer @ neurospace
 Makridakis, S., 2017. The forthcoming Artificial Intelligence (AI) revolution: Its impact on society and firms. Elsevier(90) p.46-60