Day Two of AIMed Europe 2019 was marked by a series of panel discussions on the impact of technology and innovations in medicine and healthcare, use of artificial intelligence (AI) in hospitals and healthcare organizations across Europe, and broader topics including perception of AI, guidelines and regulations, data, policy and security, and so on. 

We think we know AI but we don’t 

“Are we ready for AI?” and “How do we get ready for AI?” were respectively asked at one of the sessions moderated by Hassan Chaudhury, Digital Health Lead, UK Department for International Trade. With that, Dr. Anushka Patchava, Expert Advisor in AI, United Nations (CEFACT) thought although many claimed that they are working with AI, very few actually understand what they are really doing. 

(Left to Right): Hassan Chauhury, Jessica Morley, Dr. Anushka Patchava, Elenora Harwich, and Dr. Stylianos Kampakis

For example, in insurance, a claim is often submitted and processed by a robot, but the industry is not able to go beyond that due to the lack of data and the inability to teach machines what is relevant. Even if machines are taught, they cannot consciously handle information the way human do. As such, augmented intelligence or technology which amplify human’s capabilities to make better decisions, is preferred. 

Elenora Harwich, Director of Research and Head of Digital & Tech Innovation of Reform agreed. She added, often, people do not question who has the authority to build an AI solution, who get to use it, and the value exchange between sectors that are sharing the eventual technology and collected data. “There is a gap between promotional claims, reports from media and what is actually happening,” Harwich commented. 

Dr. Stylianos Kampakis, Research Fellow, University College London’s Centre for Blockchain Technologies said the confusion is aggravated as the general public tend to hold onto a defensive view towards AI; thinking big companies are stealing their personal data and they are not comfortable with the fact that machines have the capabilities to make decisions. 

So, how can we get ready for AI? 

Jessica Morley, AI Subject Matter Expert of NHSx suggested going back to basics; ensuring public education and the right framework and standards are in place. Dr. Kampakis echoed that, he said the government should take the lead in such an effort. Also, there is a need to invest more resources to address the Blackbox challenge or why AI is deriving at the answer as it should. 

Panel members discussing some of the questions posted by attendees sent via the AIMed Europe application

On the other hand, Harwich believed improving the accessibility of data may make a difference. Presently, there is an insufficient amount of good quality data and infrastructure which support scaling of AI-driven tools. Most of the responsibilities around controlling data, adhering to regulatory processes, and minding biases in data are all on the shoulders of entrepreneurs who are developing the solutions. Sadly, there aren’t many who will check on the quality of these training, validating and testing data. 

Dr. Patchava advocated a shift in mindset. AI is not a new way of making revenue, those who are interested should begin by understanding a problem one is trying to solve and determine how AI fit in to solve the problem. Such thinking can also be translated into an understanding of a gap between users of different demographics. She cited senior users tend to worry about the lack of human touch coming from AI, whereby the younger generation actually prefer such detach, in exchange for more privacy. 

AIMed Europe 2019 continues tomorrow (19 September). More information can be found on AIMed Europe 2019 official site, or follow us on Twitter, Facebook, Instagram for the latest event updates.

Author Bio
synthetic gene empathy chinese artificial intelligence data medicine healthcare ai

Hazel Tang

A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.