Securing the Future of Robotics

26 July 2008 at 19:55 | Posted in Ideas, Robotics, Robots, Security, Think-tank | Leave a comment
Tags: , , , , ,

[Category: Ideas. If you are new to my blog please read the “About itimes3” page first]

The 20th century started with a few early cars and planes which were very limited in performace and highly unreliable, and ended with cars and planes being ubiquitous, reliable and at an advanced stage of development.

In the same way, the 21st century started with relatively few robots with very limited functionality, and believe me, at its end robots will be ubiquitous, reliable, and at an advanced stage of development. Cars, buses, boats, planes will be robots – the human driver will be gone. The armies of the major world powers will be largely robotic. Surgeons, dentists and many other professions which rely on precision and relatively static knowledge will be robotized. Households and offices will have large numbers of robots deployed in different shapes and with different and complex functionality and intelligence. In addition, small robots will roam inside our bodies to control functions, monitor us and enchance us. And much more.

This pattern has already started in a limited way. Within a few decades, robots will start to control and manage the more mundane – and much of the advanced – aspects of daily life to a degree perhaps almost inconceivable today.

In other words: whilst the 20th century was the century of cars and airplanes, the 21st century will most definitely be the century of robotics.

I am all for robots and robotics, however there is a problem looming large, and from what I can see, lessons of the past are not being applied: this is the problem of security.

In the past two decades, personal computers have taken over much of our business lives and quite a bit of our private lives too in many cases, and as the usage of computers increases, so have threats, which keep increasing in scale and severity. Recently uncovered security risks in the DNS infrastructure of the Internet is just one of many issues that show how everything is connected and how much of the infrastructure is vulnerable.

These vulnerabilities exists because people who develop the technologies first and foremost focus on functionality – the bells and whistles. It is natural and human to do so. Security is invisible until things go wrong, and therefore is not sexy. Managers and shareholders who do not understand the technology and the risks (or who don’t care), want the product to market and just care for the functionality of it. Security is an afterthought, and will often only be addressed once a serious problem is uncovered.

The problem with robotics is the technology will interfere with our lives much deeper and on a much larger scale. Essentially, whilst computers have largely remained at the periphery of our lives (albeit connected to us and within society as a whole), robots and robotics will eventually almost fully integrate with our lives and society. They will be part of us, to such a degree that there will be a time, not too far into the future, where we will not be able to live without them.

If all computers in the world failed today economies would collapse and humanity would suffer major problems, but we would survive after some adjustments.

If all robots in the world fail 80 or 100 years from now, much of humanity will not survive because the robots will be running our lives and civilization to such a large degree.

An issue with robots is that many people still see them as individual entities: some cute humanoid robot getting us a beer or a robotic pet dog doing funny tricks. The truth is, robots will be part of an integrated “universe” which will be centrally controlled to a large degree. Robots will interact and communicate with us and with each other, and thus will be able to send and receive transmissions of data. They will need to be updated and upgraded, which will likely be a fully automated process (think an advanced version of Windows Update). Manufacturers, governments, service providers and many others will all have a measure of control over armies of robots, which implies an ability to operate, control, update them as part of a network.

Nowadays, viruses, worms, internet hacks etc. spread around the internet in seconds. The result may be that our computer or network is disabled for some time. Once the true age of robotics is upon us, this type of threat will be much more severe because complications can easily be devastating.

The artificial intelligence which many robots will possess is an added risk factor: a human programmer could implant rogue behaviour which could be hidden in the increased complexity of the software and be triggered by random events, sometimes years after the robot is first deployed. In advanced stages, a robot or robotic control entity (“central command” of a group of robots) could “seed” rogue behaviour into other robots without human intervention, and manipulate it so that it becomes vitually undetectable.

Robots will have much more impact, so instead of a virus disabling a personal computer, the “virus” that hits a robot may spur the device (or group of devices) into killing or destruction. And whilst it is inconveivable that your Roomba will rear up from the floor and headbutt you in anger for not emptying its dust tray, the much more advanced robots of the future may very well be programmed to go rogue for whatever reason: by “hackers”, by industry insiders for reasons of personal grievance, by enemies, by criminals, by terrorists, or simply because of human error during the development process of the robot “brain”.

To avoid this type of scenario, I would like to float the idea of setting up a security think-tank for the robotics industry globally, in which top-notch scientists and programmers would participate to design, from scratch ideally, secure operating systems for robots, communications and behavioral protocols, physical security and access to the software, etc. This would have to be a global effort, rising above politics and commercial interests, for the benefit of humanity as a whole (however once secure technologies have been developed and agreed upon, these can then obviously be used in commercial and individual government applications).

Rather than adapting your existing version of Linux or Windows for robotics use, I believe it is essential that new, highly secure methods be developed to create and protect the core intelligence of the robots of this century and beyond.

There will always be rogue elements trying to disrupt daily life, but when it comes to robotics if for once we don’t start with the “outside” (functionality), but rather with the “inside” (that what ultimately powers and protects the functionality) we will be much better off in the long term.

If you like this idea and you work in a type of industry where this is relevant, I would be happy to discuss in more detail, answer questions or assist in other ways. For details and contact information please see the “About itimes3” page.

George Spark

Disclaimer: Any trademarks mentioned herein are the property of their respective owners.
All usage of this site is entirely at users risk.

Blog at WordPress.com.
Entries and comments feeds.