Spying through the microwave: intelligence gathering in an online world

backPrinter-friendly versionSend to friend


Donald Trump’s advisor and spokesperson, Kellyanne Conway, raised eyebrows in March 2017 when she essentially accused former US President Barack Obama of spying on Donald Trump through a microwave. She pointed out that she did not have any evidence, but she mentioned that “there are many ways to surveil each other now, unfortunately.” Conway then delivered the punchline, suggesting US intelligence agencies might have used “microwaves that turn into cameras”. Of course, Conway says a lot of things, and it is doubtful that her claim is a reflection of a thorough understanding of the latest developments in digital intelligence gathering, but perhaps on this occasion she was not as ludicrously off the mark as it may seem.

While it is currently impossible to turn microwaves into cameras, it is true that more and more objects that we use every day are being connected to the internet. We are moving ever closer to the so-called Internet of Things (IoT), the state in which many objects will generate data and store them in a cloud. It is already fairly normal to have a thermostat with an internet connection, to say nothing of printers, smartphones and watches, and the trend towards the IoT will probably not end here. Some two years ago, one of Samsung’s CEOs predicted that in five years all his company’s products would be part of the IoT. Assuming Samsung is not the only company with ambitions along these lines, it is likely that our refrigerators, air conditioners and, yes, microwaves will no longer function offline. The same goes for our roads, buildings and even our clothes. The first toothbrushes with Bluetooth connectivity are already available for purchase.

Unsurprisingly, the trend towards the IoT has not escaped the attention of intelligence services. During a Senate hearing in 2016, then Director of National Intelligence James Clapper first admitted the intelligence community’s eagerness to exploit the possibilities of the IoT: “In the future, intelligence services might use [internet of things] for identification, surveillance, monitoring, location tracking, and targeting for recruitment, or to gain access to networks or user credentials.”

From the perspective of intelligence services, the potential of the IoT is easy to understand. Today, we can still find comfort in the things that leave no digital traces, but once the IoT takes off, pretty much every move we make will be recorded in one or more databases. This will make it much easier to find the perpetrator of a terrorist attack, for example, and to identify terror networks. Moreover, the abundance of information about our behavior also offers a broader base for predictive profiling. Intelligence agencies will certainly be interested if a drastic increase in the amounts and types of data will help them step up their game. With the introduction of its “predictive purchasing”, Amazon has already demonstrated how much the right data can tell us about what a person is going to do next.

There are, however, reasons why the IoT may not immediately be the goldmine that intelligence agencies are hoping for. So far, the big success story of predictive profiling concerns financial fraud, as it lends itself to big data analysis: it happens all the time and the perpetrators go about their crimes in more or less the same way. Therefore, the IoT could generate the data necessary to predict burglaries as they occur often enough to discern pre-crime patterns. Instead, some of the events intelligence agencies are trying to prevent, such as terror attacks, do not occur very frequently and show much more diversity. Some attacks are preceded by long radicalization processes, others are committed in the spur of the moment; some perpetrators act in groups, others alone; some terrorists undergo training, others do not. And then there are the differences (including those over time) in attack location, victim type and weapon use. Even with more types of data available, it remains to be seen whether intelligence agencies will be able to accurately predict such rare and dissimilar events.

And then there is, of course, the question of whether intelligence agencies will be able to handle the vast amounts of data that will become available as a result of the IoT. There are signs that intelligence agencies like the US National Security Agency (which houses the Central Intelligence Agency (CIA) and the Federal Bureau of Investigation (FBI)) and the UK’s Government Communications Headquarters are already having difficulties processing the data they are currently collecting. Leaked internal reports from some years ago made mention of “analysis paralysis” as a result of an information overload and claimed that there is “an imbalance between collection and exploitation capabilities, resulting in a failure to make effective use of some of the intelligence collected today”.

Admittedly, artificial intelligence (AI) is being developed to find the needles in the vastly expanding haystack; robots, for instance, will recognize faces in video footage much more efficiently than human analysts can. In fact, the CIA is currently running no fewer than 137 pilot projects to assess the potential of AI in taking on the sea of data. How long it will take before these programmers bear any fruit, is an open question, and no one knows what other types of data will be available by the time intelligence agencies have mastered the IoT-data.

It is perfectly understandable that people are concerned about the privacy implications of internet-connected devices, especially since these are generally poorly protected against intrusion from outsiders, be they hackers or government agencies. But there is a silver lining: if privacy advocates want to move on this, now is the time. Generally, governments need some time to catch up with ICT-developments, and given the complications mentioned above, things are not going to be different this time around.

Read also:

A Blockchain-Based Energy Future?
David Livingston

Knowing the nation: data and the need for vision
Victor Broers