My social media timeline is full of a Gizmodo story about a plan to use personal data, including data from Fitbits, to stop mass shootings in the USA. It is a terrible idea but it is interesting to think about some of the ways in which it is terrible, where it came from, and what’s next.
What is it?
Following the story back to the original Washington Post article the idea seems to be that a $40m-$60m research project would encourage individuals to consent to the use of personal data by a new research organisation called HARPA. The personal data would come from a range of sources, including Apple Watches, Fitbits, Amazon Echo and Google Home. At HARPA a team would analyse the data to come up with a model that would “identify risk factors when it comes to mental health that could indicate violent behavior”. The story says that HARPA will need “real-time data analytics” to stop the mass shootings.
Why is it terrible?
Here are just a few of the reasons why it is a terrible idea:
- the project assumes that targetting the behaviour of individuals is the way to reduce mass shootings, rather than other interventions like reducing the availability of guns and bullets
- the project is based on a false idea that the behaviour of individuals who commit mass shootings in the US is primarily linked to mental health. As a recent US National Council for Behavioural Health report shows the real reasons are far more complex
- the project will not generate a good model. As Emma Fridel is quoted as saying in the Gizmodo article “literally any risk factor identified for mass shooters will result in millions of false positives”. Improving the model will require the collection of ever more data about ever more people (from x% accurate, to x+1% accurate, to x+1.1% accurate accurate, etcetera etcetera while people’s behaviour continue to change). Even then it will inevitably face what Julia Powles and Helen Nissenbaum call the seductive diversion of solving bias
- the consent model is naive. Individual consent is a model that is already being challenged on the grounds of both whether individuals can ever make truly informed decisions given the growing number of use cases where data is used, and how the decisions of individuals impact on the rights of groups of people. For example, data sources like Amazon Echo and Google Home do not only collect data on the single individual who controls the account for the device but also on every individual who goes into the physical place where these devices are collecting data
- to deliver “real-time data analytics” will require data about the behaviour of individuals to be collected on a massive scale, will every individual with suspected mental health issues have data about them captured and analysed? how do we identify that group of people? perhaps just capture the data for every person in the USA?
- even with the best will and capability in the world this massive collection, sharing and use of data will create a whole host of risks and unintended consequences whether it happens in a liberal democracy or an authoritarian state
- even if an organisation could make this project work in a safe way then I would not support such mass surveillance based on my own moral values and fears of how people and societies will react to feeling like they are constantly being watched, it leads us to the data wasteland or worse
Why is the idea in the news?
The story is based on an idea pushed by an organisation called the Suzanne Wright Foundation.
This organisation was founded by Bob Wright after his wife, Suzanne, died from pancreatic cancer. One of the main goals of the organisation is to create HARPA, which would be based on the defense research and innovation agency, DARPA, but instead focus on public health issues.
It is clear Bob Wright and the foundation are well-connected in Washington and savvy enough to connect research proposals to political topics, like the mass shootings that are sadly so prevalent in the USA. There are tales that Bob Wright and the USA President, Donald Trump, know each other personally.
Unfortunately Donald Trump has the dangerous mix of embracing conflicts of interest, latching onto ideas for political gain, and wielding a lot of power.
That this particular HARPA proposal is not published openly, either by the USA government or by the Suzanne Wright Foundation, reflects badly on both of them. It makes it hard to scrutinise. It is hard to tell if these organisations really think that this idea is useful, or if they are simply using it for short-term political gains, to head off the risk of measures to reduce access to guns and bullets, or simply to create momentum for the creation of HARPA. But, it is clear that some people are concerned enough about this proposal to leak it to the press.
That should worry both the people and the organisations who might be harmed by such a terrible idea. This type of mass collection of data might seem fanciful in many countries but the USA is already seeing Amazon’s Ring service encouraging people to share data from security cameras looking out from their homes with organisations like the police.
Neither Amazon Ring’s data sharing or the Suzanne Wright Foundation’s research plan for mass shooting are likely to be effective in reducing crime, but they will both be effective at wasting money and risking unintended consequences & harm for many people. This is a shame as a government agency with both policy and delivery capability that was focussed on working out how to improve public health using modern technology, techniques and data could actually be useful.
If we want to enjoy the benefits of modern technology then the real challenge is how do we stop such terrible ideas much earlier, and well before they become horrible, horrible reality.