A Quantum Data Map for Intelligence 2025: "Follow the Data"
18th July 2020 - SISS
In recent years there have been substantial, notable improvements in the processing capabilities of computing and machine learning AI. The problem, up until last year, was that we had reached an AI rut in some respects – stuck in the vastly improved, but relatively one-dimensional Narrow AI realm i.e. pretty much all of where AI is today. Narrow AI might be described as a broadly goal-oriented approach, performing singular tasks for example, facial recognition, Robotics, Driverless Cars and so on. In essence, it is not doing anything humans cannot already do. It is automating processes.
However, the pace of technological growth has surpassed those of previous models. Moore’s Law springs to mind – a theory which, put simply, stated that the speed and processing ability of computers will double every two years and get increasingly smaller in size. Exponential Digitised Technological growth i.e. a move towards a non-linear growth pattern supporting various enterprises around the idea of the 6 D’s, meant that we saw the rate of change increasing, albeit more sporadically. But why is Quantum Computing so important? Up until last year atleast, the unversally accepted "Church-Turing Thesis" remained intact. Broadly speaking, it says that a classical computer can replicate (or process) any "reasonable" model of computation. However, when Google revealed it had entered the world of Quantum Supremacy with all guns blazing and revealed the Sycamore Processor, it became the first model that cannot be "reasonably" replicated using these classical means. In essence, it blew the lid off previous processors, including those from competitors such as Fujitsu or IBM. It has opened up a new world where the theoretical holy grail of Artificial "Super Intelligence" could happen, and not only that, could happen much faster. However, where are are we now? Even Google is cautious about making great promises and has gone to some lengths to down play the significance of this "first step" on the path to Quantum Supremacy. It is realistic in saying that Noisy Intermediate-Scale Quantum (NISQ) technology is the first marker to aim for because there are practical limitations to the100 q-bit computer as it stands today. It is still going to surpass the power and ability of classical computing, but there is a little way to go yet. That said, Google have done a good job a 'dumbing down' the science for lesser mortals and has been remarkably benevolent, but one still feels this information is being spoon fed to us. It is likely this is by design.
The next step on the AI journey and the primary target for most researchers within that field is Artificial General Intelligence or “Deep” AI. A point where machine learning does exactly that. It becomes self-learning, to an extent. There are some significant differences in opinion on this because the difficulty lies in agreeing where the point of accentuated simulation from human input meets an almost conscious, thinking system. Dystopian fears aside, this is an incredibly exciting area of course and the implications will be significant in terms of our role in the workplace and indeed most areas of society. We posted various articles over the last twelve months discussing the likely impact of IOT, 5G and the development of AI supported systems such as those used in farming, driverless cars and even sub-cutaneous chip technology (ref: “Psychopathy and the Amygdala, 2018 - SISS), which incidentally is now already about to be rolled out via the Neuralink project of the Elon Musk empire. Since writing those articles, no more than 18 months ago, these predictions and developments are already here. So, it would be a brave person who suggests AGI will not happen for a few years. Our first robot citizen of the world Sophia may not be as cutting edge as the developers want us to believe, but she (or whatever pronoun it will one day choose to identify itself with) is the pin-up bot, a marketing flagship for what is about to come. AGI is already here and as with nearly all new cutting-edge technology it is in the hands of the few for the time being. So, how does this play out for us and to the topic in hand, the intelligence services? Going back to our point about previous articles, the driverless car was a good example. We put forward the likely scenario in our report “Taxi? What’s happened to the Taxi driver?, 2019 - SISS”, that one real-life short-term victim will be the taxi driver when driverless cars become mainstream. What will happen to that person who up until then has provided a service, paid taxes, brought up a family, used public services and ‘consumed’ to the benefit of the overall economy? The likely consequences of AGI will be a huge increase in free time and opportunities to explore new avenues. Chess legend Gary Kasparov echoed our view recently, saying that he felt AI of this caliber was a blessing not a curse, and would provide opportunity to develop and create. After all that is what we do as a species, amongst other things.
This change in our roles as humans will create a seismic shift in the redistribution of resources. When that happens, and the big question is when, we will certainly be at a stage where the vast majority of tasks that could have been automated, will be automated. To then take it to the next stage i.e. more advanced AI (AGI) then processes will become exactly that, simply processes. Rather that resulting in a world where the taxi driver, the GP, the lawyer, engineer, and many public servants might spend their time roaming the earth for something to do, we will of course divert our resources into new areas. The majority of those areas are likely to involve more creative disciplines and those requiring more abstract possibly less scientific based solutions. A rational, logical question is not IF this will happen, it is WHEN will it happen, and so preparation is the key. Rather than a future to be feared (as most change is), it should be perceived as a challenge and opportunity. One key to successfully harnessing that potential will have to be an increasinlgy wide and multidisciplinary approach.
Within the world of intelligence, there are many circumstances where intelligence gathering produces new systems and innovative solutions, often on the back of military research and privately funded research organisations. Eventually a portion is rolled out to the masses after having been filtered first. We have mapped this out in two charts, one representing current data flow and the other, how we predict the data flow will change in the coming years. What it shows is that currently at the center of the data map are the Government's and their “advisors” or “data vendors”, Facebook, Amazon, Twitter or in the East, Yandex or Alibaba, Baidu and others of course. The political structures may well change and controlling factions may differ in power, but the flows are essentially the same. In the very short term, the map then points to a change, where at the center of the data flow is the “individual”. If data is the most influential, valuable commodity in the world, the direction that data travels in and what transports it to its destination will become almost as valuable. At the moment that means of transport is the internet and currently the latest ‘expansion pack’ 5G. These are the trains that carry the cargo. Other systems run along side that, such as Blockchain, Cyber Security, International Payments Systems, even CBDC and others, and these are the security guards on the train, protecting the cargo. For the time being, it might also be useful to view AI and Quantum computing (QAI) as the fuel that controls the speed of the train. In terms of Quantum AI, which way the track goes and how many tracks the data can travel on simultaneously, is a very different and difficult question to answer.
What these show is that if we can expect a power shift, a de-centralized move away from Government or the multi-national social media giants, toward the individual as a “data powerhouse”, then the role of the intelligence services will shift too. All these entities will still exist of course, but their roles in the relationship will change. That is something organisation’s like SIS will be, or should be, preparing for. The moves we can expect, will quickly go well beyond science and technology. So far, we have become accustomed to those disciplines being the drivers of change and innovation in areas such as AI – and they have been. But the creators will also eventually be replaceable, just like Facebook, just like Yandex and Baidu. Google saw this coming many moons ago and shifted its resources towards the pursuit of Quantum Supremacy and the breakthrough last year of the Sycamore Processor has established it as being well ahead of any competition, Fujitsu and IBM included. As the reverse engineering teams and developers of their own quantum machines tick away in the background in China and some European agencies, the focus may well be diverted away from where the data is going. Taking the eye off that ball would be costly, although it is highly unlikely the UK with its resources and vast network comprised of select thinkers, will not have this one thought out by now.
Onto the specifics, and the practical implications for Intelligence services, a shift in the power balance towards the individual will require an equal shift in the focus of intelligence gathering. Take for example a Government or a large multi-national, those organisation’s are populated by people. However, on the whole, when they leave the confines of that organization, the data remains. There are exceptions of course. If we then look at the multitude of cyber attacks on these groups across the world, they all tend to share one thing in common. There is one overall target. Of course, within many of the highly sensitive organizations it is not that simple, as complex networks are in place to never render a whole organization susceptible – that’s the theory at least. However, as advanced technologies come the fore and when the symbiotic bridge between AI machine and human is mainstream, what then? It is fair to say that generally when an organization or group of people in whatever entity has control of the resources, they are powerful. Without wanting to add more water to a dystopian bush that is already well out of control, or indeed to wander down the path of sci-fi such as the Matrix world, the women of the world are not just the incubators of children but incubators of data. Where we are on that blurry line of singularity at any given point in time is never possible to identify, but if we are heading past artificial general intelligence and into the realms of superintelligence, then protecting that commodity is paramount.
The intelligence services will inevitably shift in terms of the make-up of their staff, the way they gather and act on intelligence and possibly even become diluted versions of themselves in such an environment. Already we have seen how power has shifted and had practical consequences as a result of increased information and greater access to data. Vigilante groups in the UK have already achieved some tremendous success in aiding Law Enforcement departments in the tracking and eventual prosecution of pedophile rings. Indeed, the Police work in cooperation with these groups. Private Intelligence Agencies have grown rapidly too as Intelligence gathering systems become mainstream. And of course, there are the social media outlets which have now empowered the individual too and made each of us a potential Rupert Murdoch. So, it has started. This is the risk to SIS and other agencies in the medium term because everyone could eventually, very easily, become their own intelligence officer. Not only that they will become their own bank, their own social media company, their own lawyer, accountant, doctor and so on. No profession or organization will be untouched by the changes Quantum technology will drive us to at much faster speeds that previously thought possible. It was always seen as prudent and responsible for advanced government's and government bodies to plan ahead, maybe as far as 30, 50 or 100 years from now in anticipation of change for our next generation. Now move the decimal place back, as exponential rates of change demand that we do.
There is one interesting twist, however. If we agree that data is one of the most, if not the most valuable of commodities, then who decides on the quality of that data. Again, we published our view on this in the articles “The Human Stock Market, 2019 - SISS”. Any precious metal or indeed any tradeable commodity has to have a fair value, and, in most cases, “grading” of some description occurs in order to help price it. Alongside that of course there are discounts made for future supply and demand depending on the asset in question. However, if we look at it purely as a piece of gold, how do you determine what carat our data is given? Aside from our original proposition which was that this could easily create a new market purely in data and the off-shoots it provides, it is also more relevant in recent times, especially with the Trump age of “Fake News”, in essence fake data or maybe the phrase “fool’s gold” might be more apt. This can be seen in our data map, which shows the relationship between the data vendors and the Government. These vendors are tolerated by the Government because they provide a useful service, but it would be a brave person to assume that organization’s such as Facebook and Twitter are not replaceable, and not only that, almost immediately. They are. The constant in the relationship is always the Government because it influences the key three branches of control a private organization has yet to do. That is inspite of the all-powerful PACS in the US and the powerful corporations in the UK and Europe. If we throw a spanner into the works of our current assertion that data and the flow of that data will determine the direction of our future, then what do we do when fake data or ‘manipulated’ data is thrown into the mix? The simple answer at the moment is, we cannot do an awful lot. Based on the current trajectory however, we will not need to test the data for its authenticity or where it came from, because we will be our own data mines.
If this is a likely scenario, and clearly, we think it is, then Intelligence Services will have to ask themselves a slightly differently framed question. Not how can we evolve – but how can we survive?