Robotics and the public: how scientists can engage

Posted by Dr Nick Hawes on 17 February 2015

As a future where humans interact regularly with autonomous robotic devices (i.e. machines that can control themselves, such as driverless cars or robot porters in hospitals or hotels), gets ever closer, there is a growing need for scientists working in my field, Robotics and Artificial Intelligence (AI), to engage with the public about their work.

As well as the usual reasons for carrying out public engagement, there are two, interrelated, reasons for us to engage with the public about our work.

  • The first is to inform people about the real (and exciting) abilities that near-future autonomous machines will have. This is necessary to confront commonly held views about robots, largely driven by science fiction, where robots are human-like in appearance and abilities (including open-ended, general-purpose intelligence), and replace these views with a better understanding of reality, where robots are special-purpose machines with clear limits (like dishwashers or tumble dryers), but with sensors and smart algorithms to allow them to cope with a wider range of operating conditions
  • The second reason roboticists must perform public engagement is to engage with both the real concerns about the shape of a future robot-supported society and ill-founded, far-fetched (but [currently topical] worries about the threat AI and robotics poses to our existence.

To do public engagement with robots, nothing beats a live demonstration. People of all ages and backgrounds are fascinated by autonomous machines, and the chance to interact with one in person, whether it's a Lego NXT, a Nao or a PR2 is often hard to pass up.

Done well, this is also a great way of communicating the limits of the technology, as people quickly make assumptions such as "can I talk to it?" or "is it watching me?" and just as quickly see that they're wrong. The difficulty in live demonstrations is to get beyond 'what' the robot is doing and talk about 'how' and 'why' it's doing it.

Robot demonstrations

In order to do a robot demonstration well, there is really only one basic rule: focus on the people you are demonstrating to, not the robot. I’ve seen far too many bad robot demos where the presenters are too preoccupied with what the robot is, or is not doing, to engage properly with their audience. If you are new to doing such demos it is always tempting to hide behind the robot (metaphorically or literally) rather than engage with people directly, but this misses the point in why you’re doing this (to communicate something to someone).

Beyond this rule, and standard practice (prepare well, practice your demo, make everything robust), I would advise ensuring you have a number of different levels, or variants, to your demo depending on your audience. We typically start with a basic demo that shows what the robot is capable of and that we can talk through with anyone. We then have the ability to reveal an additional level of detail to explain the robot is doing what it’s doing (often this means showing visualisations of internal processing) for people who engage more closely. We also tend to also have a more interactive demo for people who just want to play or have a fun experience, without the detail (this is often allowing them to control the robot in some way).

Robot team challenges

Robot challenges, such as the DARPA Grand Challenges and competitions, such as euRathlon and RockIn scale up the idea of a live demonstration to multiple teams competing to solve a common problem (e.g. from answering a door to urban driving). Such events often provide a broader context in which to discuss robotics, and the increased separation between robot and public can provide better opportunities for communicating more complex topics, whilst the persistence of competition websites and promotional material mean that the work done for such events can have a longer impact than a short demonstration.

Online engagement with a worldwide audience

As part of the STRANDS project we recently ran an event that was part competition, part demonstration. The 2014 STRANDS Robot Marathon pitted 7 robots against each other to see which one could achieve longest, autonomous run time in a human-occupied space, as part of European Robotics Week. Some robots operated in the safety of labs, but at the University of Birmingham we deployed Bob in our very busy main library, and the winner of the 2013 Robot Marathon - University of Lincoln's Linda - went even more public in The Collection, a local museum. Wherever the robots ran they generated a lot of local buzz and thus a lot of interaction. At the same time, the online elements of the competition (leaderboards, robots tweeting autonomously, and, in two cases, online images from sensors) meant that people around the world could also engage with our work. The Marathon page received over 2,000 views from almost 800 people from more than 30 countries, with over 50% of those viewers coming from outside the UK and almost 10 per cent outside the EU.

Bob the robot's Twitter account

Whatever you are doing for public engagement, adding an online element can add an extra dimension to your work, but it is not the low-effort option people often think; the online world is now so busy that it is hard to get beyond your existing audience, but even to them, an online element can keep up interactions long after the event itself - Bob is still receiving tweets long after he completed the Marathon.

Author

In the following table, contact information relevant to the page. The first column is for visual reference only. Data is in the right column.

Dr Nick Hawes
Name: Dr Nick Hawes
Organisation: University of Birmingham

Comments

Post a comment

* Required field