On the 29 October The Samaritans launched a new Twitter app called Samaritans Radar. A search on their promotional hashtag #SamaritansRadar reveals extensive conversation, much of which expresses disquiet from people with mental health difficulties, about the surveillance function of this app.
A plethora of blog posts have offered varied perspectives and insights into concerns related to the app. Firstly @MarkOneinFour summarises the key issues being discussed on the PsychCentral blog and @bainesy1969 raises serious concerns over data privacy which issues he argues are being breached by the app. @BipolarBlogger raises ethical issues from a personal perspective in her post Mr Sam and his magical radar booth and @dr_know shares her thoughts as a mental health researcher on her blog. From a more technical perspective @akrasodomski questions whether algorithms can realistically turn tweets in to meaningful data about suicidal ideas and @adrianshort considers ethical issues related to data analysis.
These are just a few of many blog posts and tweets that reflect expert opinion and debate on the topic. An underpinning theme is one endemic to social networking sites – context collapse an idea coined by Danah Boyd (2014) – imagined and actual audiences as well as contexts collide where meaning and intentions can be easily misunderstood. The potential for well-meaning followers to take action or say the wrong thing on the basis of a tweet found through an algorithm is significant. The potential for non-well-meaning followers to add to people’s distress has also been highlighted.
I’d like to contribute to the conversation by sharing three simple reflections from a user-centred design point of view – something we are thinking about a lot in the mHealthHabitat programme which is developing the use of digital tools in health.
The allure of digital and the discomfort of unintended consequences – in my work on digital in health I’ve been struck by the fact that virtually everyone appears to have an idea for an app. There is often a misconception that an app can solve every problem and I’ve previously blogged about the allure of shiny tech here. A mobile app seems to be de rigueur for every national charity. The challenge of attempting to develop an app that can understand something as nuanced as suicidal feelings is wincingly captured in this tweet:
The concerns and worries raised on Twitter by people who experience suicidal feelings strongly suggest that a surveillance app may not the right solution to the problem The Samaritans are trying to solve. Its unintended consequences appear to be that some people may be less likely to share suicidal feelings on Twitter than they have done previously. The reverse outcome to the one intended by the charity.
Life in beta – the development of Samaritans Radar has had academic and user input to the design process as described here by Professor Scourfield. However, the conversations on Twitter and in blog posts offer rich insights and varied perspectives that don’t appear to have been considered or anticipated by those involved in developing the app. We will only know if views expressed on Twitter are a-typical if The Samaritans share their user design process and insights. This is just one constructive contribution on Twitter:
Would an opt-in system address most of the concerns around the #SamaritansRadar? Would limit number it could help but remove the intrusiveness @psychiatrySHO
I can’t help but think that this conversation would have been so much more productive in the alpha and beta stages of development, rather at the launch stage of a finished product. As a well-regarded charity, I have no doubt many people would have enthusiastically supported The Samaritans in this process. Instead they are in the unenviable position of managing vast amounts of angry feedback. It is so much harder to re-think a service or product when hours and months have already been committed to its development. Twitter is the perfect platform for distributed user-testing at key stages in development – this is an important learning point for anyone developing a mobile app.
Listening and responding – social networking sites are first and foremost interpersonal spaces where people connect, share and build networks. It isn’t always a comfortable place for institutions when public opinion is not in their favour. It is an easier thing to promote your product or service than to engage with people when things go wrong. Since experiencing criticism, The Samaritans have produced a series of short YouTube clips and an update post addressing criticisms of the app and they have amended the whitelist function. However, my sense is that they haven’t yet fully engaged with the specific issues raised and they need to do this in a more detailed and precise manner than they have done so far. If they don’t then their responses are likely to be interpreted as defensive or complacent.
So @Samaritans are intransigent: their email says what they’re doing is legal so will continue. Never mind the pain caused. #SamaritansRadar @Sectioned_
Lastly and most importantly – many people who experience suicidal feelings have articulated significant concerns about this app and for some it has created worry and distress. Above and beyond anything else, this is the issue that The Samaritans need to be most concerned about and properly address. The app must not harm those it is intended to help.