Data trackers monitor
your life so they can nudge you
Once you
know everything about a person, you can influence their behaviour. A thousand
students with tattletale phones are going to find out how easy that is
THERE'S something
strange about this year's undergraduate class at the Technical University of
Denmark – they all have exactly the same kind of phone.
The phones are tracking everywhere the
students go, who they meet and when, and every text they send. Around 1000
students are volunteers in the largest-ever experiment of its kind, one that
could change our understanding of how we interact in groups.
Sune Lehmann and Arek
Stopczynski of DTU are using the data to build a model of the social network
the students live in – who talks to who, where groups gather. They plan to test
whether the results can be used for purposes like boosting student achievement,
or even improving mental health. "We hope to be able to figure out how to
make this work in terms of academic performance," says Lehmann.
This is sociology on
a different scale, gathering detailed data about an entire group and then using
that information to "nudge" them
into changing their behaviour. Used ethically, the results could improve the
way society works, transforming everything from healthcare and public transport
to education and governance. Used for the wrong reasons, it could be extremely
dangerous.
The precedent for the
DTU experiment comes from the MIT Media Lab,
where Alex Pentland has studied the connection between human dynamics and
technology since the 90s. Since then he has run "living labs" to show
the power our social ties wield in shaping our lives. In a 2010 study,
participants were encouraged to boost their activity levels either through
personal rewards, or rewards given to a buddy who was supposed to keep tabs on
them. Being motivated by an incentivised buddy resulted in twice the activity
increase of the direct reward.
The DTU experiment is
notable not just for its size, but because it is the first time that a
community has been collectively monitored. "You get to see pretty much all
of the social ties, and the kids grow up and change a lot," says Pentland.
"They form friendships. There's a huge amount of social change. You also
know the rhythms of life, so you know when the nudge will be effective."
And what kind of
nudge should it be? Pelle Guldborg Hansen, a behavioural psychologist at
Roskilde University in Denmark, warns that it will be a challenge to nudge a
person in a way that is neither annoying nor damaging. "Creating behaviour
change would have to work with the grain of the actions being performed,"
he says, citing Google Now's subtle reminder that you need to leave now to get
to work on time as the way things should work, as opposed to
"Clippy", the annoying digital assistant built into older versions of
Microsoft Word.
Similarly, nudges
related to public health could be as simple as allowing doctors to ring up
their patients when their activity levels start to follow patterns that
correlate with, say, diabetes or depression, and asking them if they are
feeling OK.
Hansen says it's
important that these nudges shouldn't feel too personal, like a reprimand for
bad behaviour, as that might deter people from accepting them. And if someone's
activity history is used for personalised nudges in areas like socialising and
education, then ethical constraints should be in place to the same standard as
medicine, he says.
Lehmann and
Stopczynski agree, but say it is too early to decide what nudges are
appropriate. In any case, the first interventions will be gentle, Stopczynski
says, focusing on helping to raise students' awareness of activities that might
help or hinder their studies. They also plan to team up with health specialists
at the university to look for ways to support first-year students through what
can be a difficult year.
The project will have
another use too, showing what can be done with the kind of detailed data sets
that Google and Facebook already have.
"This is already
happening – Google and Facebook clearly realise the value of the data they are
collecting," Lehmann says. "We want everyone to be aware of what can
be done with the phone when the big companies start nudging you, because in the
hands of someone who wants to use this for monetary gain, I'm not sure it's
such a good idea."
If influencing
behaviour in an automated fashion en masse becomes possible, then there will
surely be unscrupulous people trying to hack the system, says Iyad Rahwan of
the Masdar Institute in Abu Dhabi, United Arab Emirates.
Good idea or not,
that world is coming. Understanding how human groups operate will give us the
tools to change the way we organise ourselves, for better or worse, Pentland
says. "If you really understand how people work, then you can build
organisations and governments that work better than the current ones do."
But we shouldn't lose
sight of the potential dark side, says Evan Selinger,
a technology ethicist at the Rochester Institute of Technology in New York.
"There is extraordinary power in the access to data at a personal level –
even predicting future behavior," he says. "There's a lot to be
gained, but there's a lot of problems that scare the living shit out of
me."
