Robot identity assurance

November 2014

Robot and human shaking handsRobots o­ffer a unique insight into what people want from technology. They also pose serious questions about ethics and regulation of identity and the internet of things. I mind about this topic in a very real sense because I expect robots to help me as I get older, says Louise Bennett Chair of the BCS Security Community of Expertise.

In the virtual world we may want to know who we are dealing with, but with varying levels of certainty according to the context of our interactions. You need to know my biological identity when you are issuing a passport, but only my avatar when I engage with you in an online game.

When I am vulnerable and a robot is helping me, a whole community of trusted individuals needs to know the identity of my robot. They will want to know what the robot is doing and be able to give it instructions that retain integrity with security, privacy and confidence.

A story that caught my eye, was about making your own robot companion by 3D printing. This was about MARC (Multi-Activated Robotic Companion). MARC’s inventor, Dr John Murray, from Lincoln, said: ‘If you are lonely, you can print yourself off a companion. How cool is that?’

At the moment MARC’s interaction has not got much beyond: ‘Pleased to meet you’, and they have only built his head and upper body, but the plan in future is to be able to download different personalities, a funny companion or an intellectual companion depending on what you are after. The inventor admits he finds it a bit creepy.

However, like Age Concern, I think there is great potential as long as we are in control. It could actually enhance our lives enormously. MARC would not be bored by repeated stories and could be programmed to laugh in the right places.

Perhaps we could have all our favourite news or comedy programmes on TV fed into our robots and they could interpret them, repeat the highlights of most interest to us and always get the punch lines for the jokes.

However, I would not want my care home manager to mistakenly download all the World Cup Football to my robot companion, when I wanted to discuss or be told about the Queen’s opening of Parliament. My learning robot would need to keep a record of my questions and our discussions to enable it to interact ever more appropriately with me, depending on my interest.

I would want to ensure those discussions and my opinions stayed private. A robot is one thing, but I would not want strangers looking into my mind.

The wrong trousers?

The second article that caught my eye was Melanie Reid’s Spinal Column in The Times. She was describing her excitement, as a paraplegic, of wearing a walking robot called Rex. The inventor Richard Little was controlling the robot trousers for her first walk.

She describes how Rex stood her up from the sofa and she was able to step forward languorously, swaying from side to side in big careful steps. She goes on to say: ‘Wallace becomes Schwarzenegger, I am invincible. I just have to relax and enjoy the view from on high again. I feel physically stretched, psychologically boosted.’

This is great, but again this shows how important it is for the right person to keep control of Rex (not a malevolent penguin) and how important it is for its internet connection to be reliable. You need absolute integrity and continuity. It doesn’t matter if there is an internet drop out on a Skype call, but it does matter if the Rex trousers freeze or fall down.

Then Paro caught my eye. Paro is one of the most often mentioned robots already doing a real world job with the elderly in Japan, Italy, Denmark, America and Australia. Paro was invented by Takanori Shibata and has been in development since 1998. He looks like a baby harp seal. Thanks to an array of sub-skin sensors, Paro responds to stroking and can turn his head at the sound of a human voice.

Paro is a comforting gentle presence on your lap or in your arms and gives the impression of following your conversation. Paro is helping in the care of dementia sufferers in a much more controlled way than real live companion animals. Paro will not be hurt if his mistress flies into a rage. Paro is easily washed and will not die. Paro makes people happy and can be a source of reassurance and calm and reduce wandering and falls.

Then I looked at a story about the Google car, which most of us have heard of. When I get too short sighted and un-coordinated to drive I will want to go out in my Google car for a spin to the shops or the seaside.

But as the headlines said: ‘What happens when it crashes?’ We are into liability issues big time here. Who is responsible? How can you be certain that my car is responding to all the inputs from other vehicles on the road? My Google car drives me along while I am on the phone or playing Angry Birds, eating my lunch or watching a film as I might on a train on dedicated tracks.

Whether you are certain of the identity of people or things, may or may not matter. Context is all. When it comes to a robot’s identity: integrity, security and privacy really matter. It is critical to have trustworthy identifiers associated with all things (just as it is for individuals) in order to know you are getting the right data to and from the right thing.

This is particularly important for robots. Are you asking the right robot to remind me to take my pills in the morning? Is the robot offering me the right pills, is it re-ordering the right pills? You actually need to know both the identity of the robot, which is directly connecting to the internet and the identity of the human being controlling the robot at any time.

What if several people can instruct the robot, say, me, my children living remotely and my doctor. Who can instruct the robot to do what actions? If the instructions conflict, who’s instructions take precedence? Who has liability if things go wrong?

So identity on the internet is both about knowing you are connecting to the right thing and knowing who is using the thing. How certain are you that it is me who is controlling this robot? There are so many ethical issues. If a robot can call on a help desk, it can communicate with other people too. This may provide a way for friends and relatives to stay in touch.

Many home automation products already offer a degree of monitoring. However, will they make adult offspring feel greater responsibility, which they might not be able to exercise, rather than giving them reassurance? Will the elderly feel snooped upon? Laws and regulation will be crucial for the way the robot market develops.

The take-up of industrial robots has always been constrained by health and safety regulations. The autonomy of lethal military robots is a serious concern that is being discussed at the Convention on certain Conventional weapons in Geneva. The spread of civilian drones for agricultural and crowd surveillance purposes, among others, will depend on the freeing up of airspace, health and safety regulations and bandwidth for their control out of line of sight.

Robots in the home will depend on liability issues as well as health and safety. Paro is rated as a consumer product in some places and is regulated in others. Denmark’s advisory Council of Ethics has looked at how acceptable it is for robots to be designed to fool people into thinking that they have feelings. The council decided that such robots were not a problem in themselves, but that carers responsible for people who might be easily fooled had to be vigilant in safeguarding the dignity of their charges.

The Alzheimer Society in the UK has decided the ethics of tagging sufferers to stop them from wandering is acceptable as it helps people to remain in their own homes. What it comes down to is that robots are as good or as bad as the people who design them and use them. As in every other sphere there are good actors and bad actors.

We need to consider the ethics and regulation of robots as part of the internet of things, before rather than after they are commonplace. There are at least four issues which have a governance dimension and that need further clarification:

  • There is a need to guarantee the uniqueness of identifiers that are linked to objects. This includes the allocation and management of IPv6 addresses (object addressing for routing purposes) and the allocation and management of RFIDTAGs (object numbering for management purposes)
  • There is a need to guarantee the security and stability of the networks that link things.
  • There is a need to avoid monopolisation of data control and to support competition among service providers.
  • There is a need to avoid the misuse of personal data (issues related to the data that emerges at the rendezvous point when an individual is linked to an object, privacy rights, the right to silence the chip, the right to be forgotten).

We also need to promote:

  • the establishment of a set of Data Protection Principles for the internet of things;
  • debate about use of big data techniques to process data harvested from the internet of things and the risks of de-anonymisation;
  • debate about the use of RFID and NFC technologies which enable tracking of individuals without their knowledge or consent.

Image: iStock/177356664

There are no comments on this item

Leave Comment

Post a comment

Blueprint for Cyber Security

Our vision is a world properly protected from cyber threat. This blueprint sets out how we can deliver that solution, starting in health and care.