Get ready for the age of sensor panic
Credit to Author: Mike Elgan| Date: Sat, 23 Feb 2019 03:00:00 -0800
A passenger on a Singapore Airlines flight this week noticed a small, circular indentation below the image playing on the seatback in-flight entertainment system in front of him. Could that be, he wondered, a camera?
The passenger did the only logical thing: He tweeted out a photo and asked the Twitterverse for opinions, setting off a chorus of complainers on Twitter.
Singapore Airlines also responded to the tweets, saying that the camera was not used by the airline to capture pictures or video. It then told media outlets in a statement that the embedded cameras “have been intended by the manufacturers for future developments. These cameras are permanently disabled on our aircraft and cannot be activated on board. We have no plans to enable or develop any features using the cameras.”
The in-flight entertainment systems are reportedly made by Panasonic Avionics, which has revealed plans in the past to use biometric passenger identification for clearing customs before you even get off the plane.
History tells us that technology deployed for convenience is often redeployed for security. Clearly AI will be capable of monitoring video from hundreds of airliner seats to find passengers who seem nervous, drunk or dangerous. Why wouldn’t airlines use the cameras for safety and security if they could? Think of the children!
Google is under fire this week for failing to disclose the existence of microphones in all Nest Guard home security systems, a product that has been on the market since 2017.
The microphone’s “coming out” happened earlier this month when Google itself announced plans to make Nest Guard voice-controlled via Google Assistant. That was when it added “microphone” to the product’s specifications list.
Google said on Twitter that until the voice control feature goes online (as the result of a software update), the microphone will have never been used.
Google also said this week that “The on-device microphone was never intended to be a secret and should have been listed in the tech specs. That was an error on our part.”
The consternation over airplane cameras and security system microphones represents a new phenomenon I call “sensor panic.” And I believe this is just the beginning.
The companies both responded to the criticism with the same basic claim: The sensors were installed for some future purpose, but to date they have been non-active.
The companies also seemed to be surprised that anyone would freak out over the existence of sensors that weren’t being used for anything.
But after what seems like daily reports about Facebook privacy transgressions, Russian hacking, Chinese industrial espionage, Android malware and all manner of leaks, hacks and privacy-invading blunders, we’ve entered into a new era of public distrust of all things technological.
For example, can Singapore Airlines and Google be trusted to not use their cameras and microphones given that they couldn’t be trusted to even disclose their existence?
(And by the way, it’s been just a couple of weeks since it was revealed that Singapore Airlines may have been secretly recording user activity on the company’s iPhone app. This harvesting of user action, which Singapore Airlines allowed without the knowledge or permission of users, was part of a service employed by the airline called “session replay” from the experience analytics firm Glassbox.)
Can these companies be trusted to prevent these sensors from being hijacked by hackers and spies?
Given the ability to process visual, audio and other data at massive scale using AI, is it reasonable to fear the sensors that gather that data?
Personally, I believe that both Singapore Airlines and Google are being honest about their sensors and can be generally trusted to not deliberately expose user data. They were both forthcoming after the complaints.
But that doesn’t even matter. We’re facing an inevitable explosion in the number of sensors harvesting data about pretty much everything, and another explosion in the availability and power of AI to process and make sense of all that data.
A startup called Xnor this week announced no-battery, solar-powered, AI-enhanced connected cameras that offer “a simple, valuable way to gather data.” The core differentiator is that the solar power and wireless connectivity mean you set them and forget them — just about anywhere.
The cameras are low resolution, and will be sold as components for other companies to build IoT devices with, rather than finished products.
The onboard AI is reportedly capable of identifying, classifying and coding objects in the view of the camera, and can then send that data to a database, programmer, control room person, spy or hacker (theoretically) using very low-data-rate and low-power IoT communication protocols. Actual photos and videos never leave the camera.
When companies build products with Xnor sensors, they’ll work instantly without being charged or explicitly connected to the internet, according to the company.
Xnor is one of several no-battery IoT startups that add up to a new trend that should worry all the people vexed by Singapore Airlines cameras or Google Next Guard microphones.
Another such company is PsiKick, which makes ultra-low-powered sensors that can harvest power through solar, movement, heat or other means. These sensors can even derive power from the very things they’re monitoring. For example, the company announced a steam valve monitoring sensor that harvests power from — you guessed it! — steam.
PsiKick guarantees that its products will work for 20 years without maintenance of any kind.
This new generation of sensors will be used for industrial purposes, in the unglamorous jobs of monitoring gauges and enabling smart meters. They’ll cut costs and boost reliability for industrial systems, factories, warehouses and power plants. Enterprises and businesses of all kinds should be ecstatic about this no-battery sensor revolution. It’s going to change everything.
And consumers should be ecstatic about another revolution. A company called Opkix this week came out with a $295 wearable video camera called the OPKIXOne that posts directly to Instagram through smartphones. Each camera is under four inches long, weighs around 12 grams, is very skinny, and records stabilized video at 1,080p and 30fps for up to 15 minutes. You can attach the camera to eyeglasses, baseball caps, jackets, backpacks, golf clubs, paper airplanes — anything, really. And you can buy two, which charge in an AirPod-like mobile charging case. These cameras aren’t in theory any more intrusive than smartphone cameras. The difference is that they can be mounted just about anywhere.
And motorists have their own reasons for excitement. Elon Musk said this week that Teslas should be self-driving by the end of the year. “I think we will be feature complete — full self-driving — this year, meaning the car will be able to find you in a parking lot, pick you up and take you all the way to your destination without an intervention, this year. I would say I am certain of that. That is not a question mark.” Also not a question mark: Every advanced new car from now on will contain an increasing number of sensors, including cameras.
All these products will be legitimate and beneficial. But they’ll also no doubt be duplicated at scale and sold cheaply online as hacker, stalker and spy tools.
Let’s zoom out and look at the big picture. We’re on the brink of an explosion in the number of cameras, microphones and other sensors in the world. Nobody knows how many, but let’s just assume that the number of sensors capable of monitoring us will increase by an order of magnitude every ten years from now on. I think that’s conservative, but you get the idea.
Our difficult task now is how to leverage the incredibly beneficial revolutions in sensor technology — and the ability to put that technology to work for us with AI — while considering the growing social phenomenon of sensor panic, and also figuring out how to assure both security and privacy in a world jam-packed with sensors.