New cities should not be built to cater
for robots and technology; instead robots must adapt to live with
humans in existing cities.
Imagine a city of the near future where robots roam freely. They
scoot along smooth footpaths, nimbly avoiding obstacles as they deliver
groceries and medicines.
They glide through supermarket aisles, detecting spills and mopping them up before anyone can slip.
They provide valuable backup support in busy, understaffed restaurants, efficiently taking food to tables of waiting diners.
Robots are already doing these tasks. However, the utopian vision of
efficient and problem-free robot helpers is a long way from reality.
Recent reports highlight some difficulties: they have had trouble
recognising people using wheelchairs on shared footpaths.
They have been labelled creepy and useless, and suspected of
following shoppers for unknown reasons. In busy restaurants, they cannot
navigate narrow aisles or customers' bags and coats.
So how can we make the best use of robots in shared urban spaces?
Cities are complex, dynamic and diverse -- this is part of the reason so many of us enjoy living in them.
But their unpredictability is what robots struggle to cope with.
While some robots can operate successfully in shared city spaces, many
can't and there is a long way to go before they'll be able.
There are several reasons for this. Most importantly, the complexity
of apparently simple tasks is an ongoing technical challenge.
Although advances are constantly being made, research by Monash
University found that there are fundamental reasons related to the logic
of their programming that shape how robots perceive the urban
environment.
The study found the growing army of robots has distinct ways of
understanding the city through what is called 'robotic logics'. Two
things are essential to their success: Predictability and connectivity.
Robots work best in situations where there are no surprises and where
connectivity to bigger data networks which robots rely on to sense,
categorise and make decisions about the world around them is available
and reliable.
One solution is to design environments that are easier for robots to
navigate. Early examples are warehouses where people and robots are kept
separate to avoid accidents. Or dedicated road lanes specifically for
driverless vehicles, as has been proposed in New South Wales.
While such responses might be great for robots, by prioritising their
needs, they risk diminishing what people value about urban environments
by simplifying them to a level that robots can manage.
Instead, the question is to ask how robots can best negotiate the
hurly-burly of our cities. Studies have shown people can tolerate
glitches or failures in the technologies they use, as they come to
understand their limits.
This means that even when people's expectations of robots are not met
because people often imagine that robots are more capable than they
really are they can still find ways to work with them.
So, restaurant workers with concerns about whether they will be
replaced by delivery robots find that a crowded dining room will
possibly literally trip up a hapless robot, so they only use the machine
when the restaurant is less crowded.
There is growing acceptance that robots perform quite simple tasks and people can step in to help when necessary.
This means thinking of robots less as replacements for human labour
and instead as partners or helpers, akin to working animals, such as
guide or cattle dogs.
A team from Monash University used an experimental scenario of a lost
child in a shopping centre to explore how people understood a robot's
limits.
They found people wanted to step into the role of decision-maker and
leave less sensitive tasks such as information about shop locations to a
robotic helper.
Another important shift in thinking is required in policy and
regulation. Policies can sometimes struggle to keep up with quickly
evolving technologies.
Experts have found that principles-driven approaches that prioritise
people's needs without quashing innovative experimental technologies are
possible.
Researchers have developed a checklist to help policymakers grapple
with the complexity of new technologies and to guide policies on robots
in public spaces.
This list links to the classic goals of public policy analysis which
include defending people and property, promoting human flourishing,
promoting efficiency and promoting social equity.
For now, robots work best when their environments are predictable --
and it is hard to assume how future cities would adapt to their needs.
This means that rather than thinking about how we can simplify cities
to accommodate robots, we need robots that work in existing urban
environments in all their diversity and complexity.
To do this we must recognise the limits of robots' capacities and
focus on how best to collaborate with them. If we can address this, then
we can support the sustainable and fair introduction of robots into our
shared urban spaces.
This is a challenge that needs to be addressed collectively by
roboticists, policymakers, end-users of robots and urban residents
alike.
(360info.org: By Shanti Sumartojo, Michael Mintrom, Dana Kulic,
Leimin Tian, Pamela Carreno-Medrano, and Robert Lundberg, Monash
University)
