There’s a lot of chatter these days about our smart, connected future, or alternately, our grim, Orwellian dystopia – the one where technology predicts what you want and delivers it before you know you want it. We’re starting to lay the foundational groundwork for that vision with the Internet of Things (IoT) and big data, but for all the noise surrounding these areas of innovation, we’ve yet to see much of anything tangible, be it exciting or harrowing.
Recently, IoT has received some mainstream attention thanks to the never-ending roll outs of basically the same fitness tracker – FuelBand, Fitbit, take your pick. Meanwhile, big data is arguably becoming the next big thing in the enterprise world, but for the average person, it hasn’t really had an impact on day-to-day life.
So what comes next? To answer that, it’s important to take stock of where we stand.
The promise of IoT is to redefine the way we interact with technology by replacing interfaces with intuitive, connected machines. The problem today is there’s an overemphasis on the connected part, with the intuitive part falling to the wayside.
Take for example LG’s Smart ThinQ refrigerator. The appliance keeps track of your food and connects to a smartphone app to let you know when it’s time to restock on orange juice, or if the milk has gone bad. While there’s certainly novelty to this, ThinQ does little to remove interfaces from our lives – in fact, it adds new ones.
A truly “smart” refrigerator would function as part of an IoT ecosystem. Rather than requiring users to manually input food information or scan bar codes, a 3D camera built into the refrigerator would recognize items automatically. The refrigerator could then interface with another 3D camera in the dining room that not only recognizes when you’re having a meal, but can delineate meatloaf distinctly as meatloaf, and eventually, that Wednesday is meatloaf day. If you’re out of ketchup come Wednesday, the refrigerator connects to your car’s GPS to infer the most convenient grocery store for you to restock at. That information then gets pushed to your smartphone.
Connecting devices in and of itself isn’t particularly difficult or valuable. The value occurs when devices and sensors can share information and draw new conclusions from the aggregate. That aggregate is what we call big data.
Like IoT, big data is framed as a shiny catch all solution, but proponents tend to overlook some basic shortcomings. A 2013 article in the Harvard Business Review makes the obvious point that good data doesn’t necessarily lead to good decisions. A study by the UN from 2013 argues that analysis driven by big data is fundamentally “informed by the world as it was in the past, or, at best, as it currently is.” That is to say big data is useful in predicting the future only inasmuch as the future resembles the past.
Still, companies like Netflix are already making great use of big data to drive everyday decision-making. Ayasdi, a firm that grew out of research done at DARPA, offers topological big data analysis to some of the world’s largest conglomerates, ranging from Citibank to GE. The company says its analysis tools can predict anything from acts of terrorism to the discovery of new petroleum resources.
Perhaps the greatest challenge with big data is that we haven’t figured out the scope of its utility yet. Sure, Netflix can leverage existing customer data to predict the behavior of new users, but can we use a similar approach to predict things like the socioeconomic effects of an earthquake or tsunami? That remains to be seen.
More importantly, the quality of any data set depends on the sensors collecting the information. This is where big data and IoT become two sides of the same coin. The more sensors and devices we have, the more comprehensive our data set becomes. More comprehensive data leads to more intelligent devices that can accurately predict desired user outcomes without user input.
Where It Might All Go Wrong
The notion of a future full of predictive, ubiquitous machines tends to evoke scenes from Minority Report, but it’s worth noting that it was actually the film’s human element many found most disturbing; not Tom Cruise’s giant hologram computer, or even the personalized ads that follow him around the city, but the precog Agatha, endlessly lost in some trippy, prophetic hellscape like the demented lovechild of Timothy Leary and the Borg Queen. How would we ever know if she could really see the future?
In the real world, we need look no farther than the city of Camden, New Jersey for a case study in tech-driven, ubiquitous surveillance. In 2012, the city had the highest crime rate in the country and then in 2013, its entire police force was disbanded due to budget cuts.
Today, a police force run under county auspices patrols the streets, aided by one of the most sophisticated surveillance systems anywhere in the U.S. A network of 121 cameras and 35 microphones monitors every city block 24 hours a day. The microphones can triangulate the sound of gunfire to within a few meters, and the cameras respond by retraining on any possible escape routes. This information all gets funneled to a state-of-the-art $4.5 million command center where valuable insights are derived from the aggregated data. This is IoT and big data in action, and it’s working.
We shouldn’t be worried about some imagined Orwellian surveillance state. The true danger in predictive tech is much less dramatic and a lot more insidious: a protracted standards war. Rather than an Internet of Things, we get several Internets of Things that are incompatible and don’t talk to each other. Instead of big data, we get several moderately sized silos. Remember the smart refrigerator? What if it only reminds you about the ketchup if you have an Android smartphone and Daimler automobile?
Fortunately, the open-source movement has made great strides in evangelizing the merit of open platforms. Startups like Spark are already working to ensure the unity of IoT through robust APIs.
“Today, there aren’t enough things on the market to worry about intercompatibility, but there will be five years from now,” Spark’s founder recently told Wired.
With any luck, we’ll have figured out how to share by then.