Robotics and the internet of things – implications for insurers
A large number of employees at a major Chinese smartphone supplier are to be replaced by robots. The company announced in December 2016 that numerous tasks would be taken over by robots made in-house, with ultimately only a few employees remaining in its factories in production, logistics, testing and inspection. The idea could become popular, as robotics is used in practically all parts of industry, and increasingly in day-to-day life. What are the implications for the insurance industry, and what challenges will it face?
Modern robots learn by themselves
Robots have been used for decades in vehicle manufacture. They enable processes to be automated – moulding, welding, painting and assembly. The benefits are obvious. Robots require little maintenance, are not sensitive to the environment and can achieve a high level of precision and efficiency. Our digital friends are also becoming increasingly common outside manufacturing. In Japan for example, robots can often be found in care homes interacting with the residents.
Using robotics merely to automate is already a thing of the past. Many modern robot systems are now based on self-learning technology. No longer do they "merely" put things together quickly and in conformity with standards. They are now also networked via the internet of things (IoT), and in future may well be taking decisions autonomously. For example, industrial robots plan the quantity of raw materials they need, thereby continuously optimising their capacity utilisation.
Risks in IoT robotics
Apart from the obvious benefits such as increased efficiency and precision, there are risks in robotics. Any facility linked to the internet is exposed to cyber security risks. Any piece of equipment with its own IP address offers cyber criminals a port of entry for the theft or manipulation of data. Moreover, software errors or misdirected links in systems create risks that can result in substantial loss of production and data, or property damage.
Whilst the risks around industrial robotics can primarily cause cyber or property damage in the immediate surroundings, i.e. to production machinery and factories, the use of robots in homes and in the public space also poses risks such as personal injury. For example, incorrect intervention by a robot used in care could cause injury to a patient. Care robots could also be involved in road accidents in which people are hurt, while accompanying patients on walks for example. In addition, risks can also arise outside industry in the area of product liability if damage is caused to third-party property.
Impact of IoT robotics on underwriting
"The latest developments must be fully reflected in risk assessment", says Marcel Koos, Head of Corporate Risk Consulting at Munich Re. "Since devices are connected to the internet, processes that previously could not be manipulated now can be. In industry in particular, a single loss event can affect a whole range of companies because the same control systems are used and facilities are connected to the internet." This can result in considerable accumulation losses. "With self-learning systems, it is important to install strict end controls for products and applications so that any "misguided learning" can be detected as soon as possible. It is also imperative to set clear limits on robots' freedom of action", emphasises Koos.
Despite the new risks and the challenges for risk assessment, it is likely that, overall, claims costs will decrease and there will be fewer manufacturing errors in many areas. "From a risk assessment point of view, the use of robots makes sense and reduces the risks, especially where work is hazardous, dirty or harmful to health", adds Koos. "And that could have a positive effect on covers such as workers' compensation and employer’s liability."
Robotics and IoT: Liability issues
Assignment of liability will not infrequently prove to be complicated. How quickly must a user react to an alarm from the robot? Whose responsibility is it to protect networked robots from hackers and what measures do they have to take? What is the situation if in industry complex robot systems networked worldwide are faced with differing – or even contradictory – national liability rules? There will often be long disputes over recourse involving a number of parties, for example if the keeper or its insurer are initially liable, but then attempt to pass the loss on to the manufacturer by proving that there was a product defect. Some of these problems could be mitigated by amendments to current law. One possibility would be a general strict keeper liability for robots similar to that existing in German law for certain household pets. Or an extension of robot manufacturers' product liability, with the presumption of a product defect if a robot causes damage. The law would have to take account of both consumer-protection aspects and economic considerations such as the need to promote new technologies.