In December, my colleague Richard Hutchings (Senior Research Manager) and I attended the BHBIA Winter Seminar in London to hear all about the topic, which was: “Let’s be Frenemies… You, Me and AI”.
As this was my first market research conference, I was excited by the opportunity to hear from a variety of perspectives about the impact that artificial intelligence and machine learning may have in the future – including the idea of AI developing ‘consciousness’ and the ability to be creative.
Having joined HRW’s innovation team about a year ago, I’ve been following the impact that big data, social listening and virtual reality has in this space and this year’s conference covered all that – and more!
Typically, the speakers began discussing AI in the wider world – whether it be robots with personalities, an AI scouting platform for premier league football, or teaching a computer to paint an image in the style of Van Gogh – clearly emphasising how AI can impact all walks of life. A running theme was how the information we can obtain and the availability of big data is exponentially increasing and is much more robust and accurate than human ‘self-reporting’ – a concept that was colourfully demonstrated by Dr Ali Goode who presented inaccurate statistics of average male genital size as a result of self-reporting.
When discussing healthcare specifically, most described AI’s applicability in diagnosis – with a few mentions of the ability to identify cancerous cells and predict tumour progression. We heard from Dr. Andree Bates about an interesting case study using a social listening/ facial recognition tool to identify facial patterns for children with a rare (but fatal) genetic condition and dropping subtle online prompts for their parents to seek medical attention.
Throughout the day there were mentions of some market research techniques that we’re familiar with at HRW such as reaction time (to mitigate issues in self-reporting – request our 2016 webinar here), augmented reality (for device testing – read our 2017 paper here), virtual reality (to help respondents make judgements in a future market place as we saw in our 2016 self-funded research) and voice ‘emotion’ processing software (to identify and quantify moods in patient research – read about our 2018 pilot).
Many of this year’s speakers ended their talk with a caveat that AI is only as good as the data you give it – highlighted with examples such as the Microsoft AI chatbot that had to be disabled after 24 hours due to Twitter users teaching it to be very very politically incorrect. While there was an air of optimism for what AI could bring to the healthcare space in the future (such as developing personalised medicine by predicting patient outcomes), we were cautioned by the fact that AI loses accountability and cannot understand the rationale behind trends so ultimately should not be given the final say.
Rich and I agreed that what we’ve seen through our own use of machine learning at HRW (evolving decision trees with advanced machine learning algorithms such as Boosted Trees, Bayes nets, and Random Forests) is that the success of these algorithms depends on human help and teaching, and that for so many clients the AI accountability problem can be a hurdle to acceptance.
Start to finish, the day was filled with stimulating conversation and the opportunity to meet and engage with a variety of other attendees – and of course a nice get together over the delicious three course meal.
I am grateful for having had the opportunity to attend this event and look forward to my next one! Hopefully I will see some, now familiar, faces.
By Francesca Cooper