IoT news of the week for March 16, 2018

This is what we all fear when talking about security and IoT: A few weeks back I wrote about a vulnerability in Schneider Electric’s controllers that was exploited and the company’s reaction. But the New York Times lays out some very scary facts around a flawed attack of a petrochemical plant in Saudi Arabia. Basically the story acts as confirmation that our worst fears around critical infrastructure and how vulnerable it can be to malicious acts by nation states. It appears that the attacks required physical access to the plant’s network, rather than being uploaded using the internet, but the fact is that as we move to digital control of devices, even those that aren’t connected to public networks are vulnerable. (NYT)

DCI Alexa at your service: Police in England are letting citizens hear about police reports and crimes committed in their area via their Amazon Echo. The next step is letting people report crimes via Alexa. There are a lot of questions about this idea, ranging from what kind of help one could expect to how soon would one get it after calling Alexa. I also wonder how it further divides people along class or even privacy-loving lines. (The Intercept)

Alexa gets a business degree? In an interview with Werner Vogels, the CTO of Amazon, he talks about plans to introduce Alexa to business environments. The article glides across complex topics such as making sure companies can write their own skills for Alexa and if their software can integrate with the smart speaker. I worry that until the challenge of matching specific voices to requests is met, Alexa will mostly be there to make conference calls run a little more smoothly vs. offering up much business insight. There are so many different layers of access and permissions in business environments that we can’t really replicate using just our voice as the key. But walking into an empty conference room and asking Alexa to book it for a future meeting would be handy. (Axios)

Nest has a new temperature sensor: Whelp, it’s about time! Nest finally has a temperature sensor that will work with its thermostats, allowing users to take temperatures from around their house, not just where the thermostat is. The sensors cost $39 for one or $99 for a three-pack. Nest also released its video doorbell options, giving it a credible (but expensive) home security system offering. (CNET)

L.L. Bean backs out of IoT plans: In February, L.L. Bean teamed up with sensor maker Loomia to put connected tags in the sportswear company’s boots and jackets. It wanted to use the sensors and the blockchain to track what happened to its gear after it was sold. The goal was to see how often it was worn and in what conditions. It has now quit the project after concerned consumers worried that L.L. Bean would be tracking everything from their location to their activities in the sensor-laden gear. I’ve said this before, and now seems like a good time to say it again: Without reliable and known privacy safeguards, the internet of things will never win over the mainstream. We need regulations, and companies should be helping to drive them forward. (WSJ)

AI researchers need to develop a security mindset: How often have you sat at a traffic light marveling that the only reason there aren’t more deaths behind the wheel of a car is that most people understand and follow arbitrary rules? Most people don’t run red lights, for example. Most people try to stay in their lane. (This isn’t necessarily true in other parts of the world, but no matter where you are, there are still rules that everyone has agreed to follow.) Unfortunately when everything’s connected we have to assume that everything is vulnerable to the few who don’t follow the rules. And that includes those designing AI for use in the real world. This article talks about how researchers are being spoofed by those who want to test or disprove the validity of AI models, and argues that when one pixel can be changed to fool a computer, designers of automated systems need to start thinking like their adversaries, much like security researchers do. (Wired)

Is data portability the answer to data monopolies? The advantage that big platforms have when it comes to bettering their products using machine learning is astonishing. Good AI depends on good data sets to train the AI, and giant tech companies have this data in droves. But in D.C., regulators are starting to ask questions about what it means that competitors can’t seem to accrue the same data advantage. One solution is to make it easier for consumers to port their extant social graphs to other platforms. In practice, such an effort would be difficult, given how this type of data is stored and how interlinked it is. Other solutions involve escrows or regulatory bodies that might oversee the tech giants. I think that as the GDPR regulations take effect in Europe, we’ll see a lot of innovation on this front, so it may make sense to wait a few more months and see how things unfold. (Forbes)

Meet the new heads of the Industrial Internet Consortium: This week, the Industrial Internet Consortium changed its leadership, bringing in four men from Dell, Huawei, Real-Time Innovations, and Bosch. The IIC launched with much fanfare in 2014 and has since worked to create frameworks for various industrial internet test cases. So instead of watching over a specific protocol, the organization helps its members pull together entire projects and certifies those as interoperable. Outside of the members, I haven’t encountered a lot of companies using IIC-certified frameworks, but let me know if you do. (BusinessWire)

Check out a new guild for DIYers: My friend, Dr. Lucy Rogers, who has appeared on the podcast, has set up a new effort for those of us who love Arduinos and playing with solder or glue. The Guild of Makers launches Friday in the UK, and I’d love to see something like this form in the U.S. as a resource for folks who are trying to figure out how to flash a Pi or knit a sweater. (Guild of Makers)

The 2018 Design in Tech report is out: John Maeda thinks on the topic of design and how we should bring design principles to everyday experiences mediated through technology. This may mean a focus on web site design, the flow of a user through a shopping site, or even how a user interacts with a connected object. A big theme in this year’s report is how we experience AI-driven choices in our tech lives. The report also represents an experiment of sorts for Maeda, who learned various coding practices in order to make the report more interactive and responsive. Sadly, it doesn’t degrade gracefully and the resulting slides can be difficult to read. This feels like a design mistake, especially when connectivity isn’t always a given. (Design in Tech)