I get it. Not everyone listens to podcasts. So to make sure you don’t miss some of the best ideas I’ve heard in consumer, enterprise, and industrial IoT, I’m compiling them in a year-end story to enable you to walk away smarter and ready to take on next year. Check out the ideas below and feel free to let me know what you think.
And feel free to share your suggestions for guests over the coming year, too.
Medical sensors in earbuds: Poppy Crum, chief scientist at Dolby Laboratories, came on the show in April for what felt like one of our more philosophical episodes. We spent a lot of time talking about how our physical attributes shape our perceptions and how technology should respect those unique attributes. She also shared some details about the human ear that surprised me. She said that in-ear medical monitoring can provide details on heart rate, brain functions, blood oxygenation — and even galvanic skin response. She may have predicted the future, as Apple has been rumored to be thinking about ways to put medical sensors in Airpods. Listen here.
Adding a data disclaimer as you put items in your shopping cart: This idea comes from Andi Wilson Thompson, a policy analyst at New America’s Open Technology Institute. She and I chatted about digital rights, privacy, and device security. We also talked about the frustration a user might feel after they’ve purchased a device and then connected it, only to be faced with onerous terms and conditions. What if they don’t accept them? Can they return the device at full value? If you bought a light bulb, yes; but what about if you buy a car? Or a connected home? Instead of telling someone after they purchase the product, she suggests sharing a data usage notification when something is placed in a shopping cart. So if you pick up a smart device on Amazon.com, perhaps you’d see a pop-up that tells you the device data is shared with third-party companies, but is anonymized. Listen here.
When it comes to predictive maintenance, find your vitals: When journalists write about the ability to use sensor data to predict when a machine might fail, it’s often presented as a binary. Their articles assume some AI will take in the sensor data and spit out a date and time when the machine will fail. But as Chris Smith, vice president of service innovation at Otis Elevator Company, explains, knowing when a machine will fail is a really hard problem to train for. Instead, Otis’ sensors and AI look for the “vitals” for the machine so his engineers can get a sense of the overall health of the elevators. From there, the algorithm can predict roughly when an elevator might fail and schedule maintenance before it happens. Otis arrived at this strategy after trying to find an algorithm that would offer precision, and failing. Instead, having a good enough set of factors to measure worked just as well, if not better. Listen here.
A privacy-friendly way to anonymize location data: After The New York Times published its exposé on apps collecting users’ detailed location data, more people became aware of how even anonymized location data can be used to easily identify them. But it doesn’t have to be that way. Emily Silverman, program manager for the City and County of Denver, told me how Denver tracks cars at intersections for traffic purposes using a randomly generated number for each five-minute segment of the trip. The point here is to focus on the number of cars in a particular place, not figure out who’s heading where. If the data aggregation is actually trying to identify consumers’ visits to various locales to create highly accurate demographic profiles, this is not a helpful strategy. But to track the popularity of places or the number of people in a particular location, this is a viable and privacy-preserving option. Listen here.
Start your digital transformation with a lot of meetings: This sounds terrible because, let’s face it, no one likes meetings. But as Entergy‘s Raiford Smith makes clear when talking about how he led that company’s digital transformation, conversations are a necessary tool for getting people on board. He suggests that the digital transformation team find problems that different business units want to solve with IoT and AI, which will help them understand how the shift will specifically help them. He also recommends being up front about which projects are realistic and will be completed first. So much of what I hear about digital transformation efforts are focused on telling people how to change or just putting new technology in place, whereas Smith’s approach is very people-centric. Listen here.