In my opinion, an old Finnish proverb applies rather well to the Internet of things (IoT). It is a good servant, but a poor master. Connected sensors and controller units serve as a bridge between the digital and the real worlds. However, it is good to keep in mind that we all live in the real world. We interact with matter using our bodies, which exists beyond the digital dimension. Efficient, direct online communication by devices may definitely help out a lot, at its best, but it must not become an end, in and of, itself.
When the IoT servant becomes the master, information security often becomes a problem. F-Secure’s research manager, Mikko Hyppönen, gave a startling presentation on the information security risks of IoT in an IoT seminar a while back. He explained how his company regularly scans the entire IPv4 address space of the Internet several times a week and how they, time after time, find thousands and thousands of devices connected directly to the Internet without the slightest attempt at implementing even password protection, not to mention a firewall. It goes without saying that it is bothersome if anyone, without authentication, can start up or shut down a power plant generator or adjust hospital bed settings, if they only happen to come across the right IP address.
Another part of the problem, and I think this is more important, is that of excess automation. The task of automation should be to make operations easier. Automation, in itself, should not become an end and it should not be “too smart.”
Almost everyone has encountered a situation where a computer stubbornly makes wrong guesses. For example, when overly active automatic correction or predictive text entry always replaces the right word with a wrong one. Or, when personalised advertising, with all good intentions, stubbornly offers advertising on products that you are not the least bit interested in, just because someone else at the same IP address as you is doing or has recently done searches related to the topic.
In the seminar I mentioned, Tesla’s Northern Europe director proudly announced how the car they manufacture automatically rises up when arriving at a rough road where another Tesla driver has asked for more road clearance. I, for one, do not want my car to work according to someone else’s wishes, but instead want to keep control of things, myself. Of course, I could accept suggestions, as long as the ultimate choice was mine.
This problem that I bring up has already been identified in automated decision-making. It is an interesting question; how much power will we give to machines or, rather, the algorithms within. Cautionary examples can be found in the investment robots of the financial world or in automated online gambling. We have already learned how automation may turn against itself or completely demolish the established laws of some activity. In principle, machine logic, at its best, can work a lot better than human reasoning as long as the algorithm is able to take all of the necessary aspects into consideration. The challenge of assessing all related causalities increases with the growing complexity of concepts.
New, “smart” IoT equipment should stick to its role as servants, just like computers have. I will never forget the half-joking words of Kari Jouko Räihä, professor at Tampere University, on how everyone should stop talking about the user-friendliness of computers. “A computer should not be a friend. It should be a servant.”
Networked automation will inevitably become more commonplace, and it will be the task of us professionals in the field to keep it in its role as a servant. We must find the right balance between automation and the related human interactions. I do not want my fridge to automatically order more beer for me without checking with me first; although, I will be glad to accept the car automation correcting sliding that was caused by my mistake.
I want to make the decision for my things now and in the coming world of networked things.
Mika Naatula, SVP, Business Solutions @ Enfo