When IoT systems fail: The risk of having bad IoT data

No matter what numbers you look at, it’s clear that the internet of things (IoT) continues to worm its way into more and more areas of personal and private life. That growth brings many benefits, but it also poses new risks. A big question is who takes responsibility when things go wrong.

Perhaps the biggest issue surrounds the use of IoT-generated data to personalize the offering and pricing of various products and services. Insurance companies have long struggled with how best to use IoT data, but last year I wrote about how IoT sensors are beginning to be used to help home insurers reduce water damage losses. And some companies are looking into the potential for insurers to bid for consumers: business based on the risks (or lack thereof) revealed by their smart-home data.

But some of the biggest progress has come in the area of automobile insurance, where many automobile insurers already let customers install tracking devices in their cars in exchange for discounts for demonstrating safe-driving habits.

The rise of usage-based insurance

Called usage-based insurance (UBI), this “pay-as-you-drive” approach tracks speed, location, and other factors to assess risk and calculate auto insurance premiums. An estimated 50 million U.S. drivers will have enrolled in UBI programs by 2020.

Not surprisingly, insurers love UBI because it helps them calculate their risks more precisely. In fact, AIG Ireland is trying to get the country to require UBI for drivers under 25. And demonstrably safe drivers are also often happy save some money. There has been pushback, of course, mostly from privacy advocates and groups who might have to pay more under this model.

What happens when something goes wrong?

But there’s another, more worrisome, potential issue: What happens when the data provided by the IoT device is wrong or gets garbled somewhere along the way? Because despite all the automation, error-checking, and so on, occasional errors inevitably slip through the cracks.

Unfortunately, this isn’t just an academic concern that might someday accidentally cost some careful drivers a few extra bucks on their insurance. It’s already a real-world problem with serious consequences. And just like the insurance industry still hasn’t figured out who should “own” data generated by customer-facing IoT devices, it’s not clear who would take responsibility for dealing with problems with that data.

Though not strictly an IoT issue, computer “glitches” allegedly led to Hertz rental cars erroneously being reported stolen and innocent renters being arrested and detained. The result? Criminal charges, years of litigation, and finger pointing. Lots and lots of finger pointing.

With that in mind, it’s easy to imagine, for example, an IoT sensor getting confused and indicating that a car was speeding even while safely under the speed limit. Think of the hassles of trying to fight that in court, or arguing with your insurance company over it.

(Of course, there’s also the flip side of this problem: Consumers may find ways to hack the data shared by their IoT devices to fraudulently qualify for lower rates or deflect blame for an incident. There’s no real plan in place to deal with that, either.)

Studying the need for government regulation

Given the potential impacts of these issues, and the apparent lack of interest in dealing with them from the many companies involved, it seems legitimate to wonder if government intervention is warranted.

That could be one motivation behind the reintroduction of the SMART (State of Modern Application, Research, and Trends of) IoT Act by Rep. Bob Latta (R-Ohio). The bill, stemming from a bipartisan IoT working group helmed by Latta and Rep. Peter Welch (D-Vt.), passed the House last fall but failed in the Senate. It would require the Commerce Department to study the state of the IoT industry and report back to the House Energy & Commerce and Senate Commerce Committee in two years. 

In a statement, Latta said, “With a projected economic impact in the trillions of dollars, we need to look at the policies, opportunities, and challenges that IoT presents. The SMART IoT Act will make it easier to understand what the government is doing on IoT policy, what it can do better, and how federal policies can impact the research and discovery of cutting-edge technologies.” 

The research is welcome, but the bill may not even pass. Even it does, with its two-year wait time, the IoT will likely evolve too fast for the government to keep up.