How AI Disrupts Healthcare: a German-Canadian Perspective


How AI Disrupts Healthcare: a German-Canadian Perspective

On March 28, 2019 the conference AI in Healthcare brought together experts from Canada and Germany to discuss how the deployment of artificial intelligence (AI) disrupts the health sector. Over 90 people attended the event at the Canadian Embassy in Berlin that was organized by the Canadian German Chamber of Industry and Commerce (CGCIC) with the help of Smart Data Forum.

“Germany and Canada are more than just dating, it’s a love affair since 1971”, said Ghislain Robichaud, Counsellor for Science and Technology at the Canadian Embassy, as he began his insightful presentation on the Canadian AI landscape as well as opportunities for cooperation. Afterwards Michelle Lau, Venture Manager for Digital Health at the Creative Destruction Lab (CDL) in Toronto, explained how CDL helps tech founders to launch their business ideas successfully, presenting startups for clinical cases (e.g. oncoustics) and for operational advancements (e.g. bridge7).

One major discussion point at the conference centered around the question of how to build a health database for effective AI usage. One key challenge identified was how to weigh between the quality and quantity of data, another was that consistent standards and protocols for data sets that are yet to be developed. Prof. Dr. Thomas Zahn, managing director at the research institute GeWINO from the insurer AOK Nordost explained that due to Germany’s public health system that covers over 70 million people, the data set in Germany is quite fruitful for studies. Quite different is it for Canada where there is a fragmentation of data between the different provinces.

Another discussion topic referred to the advanced pattern recognition through AI systems. When discussing the challenges in adapting AI in healthcare, Dietmar Frey, project leader of PREDICTioN 2020, raised the point that AI startups have to show that their tools provide better outcomes and benefits for the patients. Kiret Dhindsa, postdoctoral fellow working on machine learning and data science, cautioned against implementing biases in AI systems. Prof. Dr. Klemens Budde, senior physician at Charité University, suggested that AI should be seen as tool – like a stethoscope – that doctors have to learn how to use well and that will bring great benefits to diagnosing.

The conference concluded with a panel discussing the disruptions resulting from easy-to-use do-it-yourself diagnostic tools. Yannick Schmid, Business Intelligence Developer at the startup Vivy, that develops patient-centric electronic health record, explained that Vivy is encrypting all patient-related data and therefore is not able to use the data for training AI. Vivy would rather empower the patients to take care of their own health. Necessary condition for the useful application is that healthcare is no longer thought of in distinct silos but rather as a cohesive ecosystem. This change would also help focusing on preventing sicknesses rather than only curing broken out diseases. Eduardo Peire, founder of AIScope, concluded that he sees AI as the perfect companion for diagnostics.