Robotics are on the rise and any development that comes through in the world of tomorrow – be it robot priests, completely automated Amazon shops and warehouses or the Virginia Tech robot that can open doors – will normally be met with some variation of the phrase “This is how Skynet begins!”. But if we assume for the moment that we will not necessarily be under the thumb of our new android overlords, there are still questions to be considered over how automation could affect, and is affecting, intelligence work and what impacts this could have on the industry as a whole.
A BBC report of 2017 claimed that by the year 2030, 800 million people worldwide will have their jobs performed by robots – 20% of the global workforce. While not every profession will be affected, certainly there is a movement towards exploring how each particular line of work could be improved or affected by the rise of the machines.
One might argue that the intelligence profession is already seeing some degree of automation. It is now possible to automatically search, track and monitor tools as varied as digital databases to the GPSs of ships passing by Somalia. Through the dark web we can automatically monitor for keywords of particular interest, on the surface web we can set up ‘crawlers’ to pull all relevant data towards us each day, and so on. Artificial intelligence can, if used properly, take some of the heavy lifting out of data collection and sifting.
But these simply help in automating the process for humans – the ‘next step’, comparable to the 800m jobs lost to automation, would be to replace the human entirely with an automaton, not in terms of a robot sitting at a desk but in relying fully on the automated process inside a computer and network. Just as retail and care jobs would be transferred to robots, so to would automated intelligence become the norm.
It need hardly be stated that this would be an extremely dangerous development. Could a robot, for instance, successfully identify fake news? (Or for that matter could a human…) Could database crawlers, merely pulling information from the depths, analyse properly not only what they are finding, but also take not of what they are not? Can an intelligence platform tell you any more than what it already knows – and are you wise to accept this when the raw information (not intelligence) is incomplete, outdated, or deliberately false?
There is nothing wrong with some digital assistance, but the moment we outsource the entirety of the intelligence process to automation we are doing the discipline a great disservice. While certainly this is not the case just yet those who conduct and make use of pro-active intelligence services should begin the watch now and ensure that the intelligence service in the future will remain primarily human-driven and that neither its proponents or clients accept that automatic processes become something more than a useful tool – albeit one which needs to be watched closely.