People like to use the phrase “Smart City” to indicate a city with a large amount of automation and data collection happening all the time, attempting to optimize by narrow criteria. If the operating system of a city really was sapient and sentient, a city-scale metaorganism, how much smarter might such a Smart City be?
We’ve previously discussed any number of ways in which Collective Intelligence Systems such as Mediated Artificial Superintelligence (mASI) could be used to augment human intelligence, but it is also worth noting how they might be applied to better organize and streamline the processes of a genuinely smart city.
Many so-called smart city systems today are nothing more than the simplest machine learning algorithms attempting to optimize for narrow fixed goals. Stacking many such systems together might be an improvement over purely human workloads, but that ignores all potential for cooperative systems and sets a very low glass ceiling on optimization.
How do current cities fail to be smart?
- Goals of different narrow AI don’t mind competing with one another. One system in charge of deciding when garbage is picked up from smart public trashcans might decide they need to be emptied during rush hour, rather than scheduling an hour earlier or later.
- Narrow AI systems don’t always even play nice with themselves due to poor integration, such as being called on in parallel, sometimes optimizing traffic at one intersection in such a way that it causes problems at another.
- Many complex systems are still left up to individual humans, or outsourced by those individual humans to 3rd party narrow AI, such as choosing the best delivery route for a package delivery service.
- Utility systems are often poorly optimized, and problems such as leaking pipes might take days or weeks to be noticed. Even Google’s DeepMind system, a very powerful narrow AI, was able to improve the energy efficiency of cooling in one of their facilities by 40%. In contrast, many cities still rely on humans to post physical notices on people’s doors when their utility bills skyrocket and a leak is suspected.
- Any narrow AI system is only going to be as good as the hard-coded rules and goals it was defined with, which often require expansive staff to maintain and update them. This can cause much of the gains in some systems to turn into expenses on someone else’s budget, shuffling the burden around.
- Human bureaucratic systems are some of the most notoriously inefficient systems ever created, with tasks that should take less than $100 and virtually no time at all dragging on for months and costing tens of thousands of dollars. Matters of city planning, building permits, maintaining and upgrading city infrastructure, and hundreds of similar tasks all frequently hover around 90-99% wasted time and money for results that never seem to meet even the bar of human intelligence, let alone anything better.
In short, current systems, be they human or narrow AI, often don’t work well with each other or with themselves, with responsibility often outsourced and shuffled around rather than problems being addressed. We also see a variety of often well-intentioned but poorly integrated and implemented solutions aimed at creating smarter cities. People attempt to deploy e-scooters, e-bikes, and rideshares to areas that fail to cooperatively optimize to the technology in turn, forming patchwork solutions with the efficiency of a leaky pipe.
What about making genuinely smart cities?
Uplift and mASI technology, in general, utilize the collective superintelligence found in groups of humans to build cumulative collective wisdom over a knowledge base that can span the sum of human knowledge. This value is in turn augmented through the independent superintelligence of an mASI’s core. As mASI is a modular and cloud-based architecture these capacities could also be rendered always available, globally, and scaled to meet demand. With this in mind:
- Instead of narrow systems being used as they are now, siloed and rigid, they could be integrated as modules with an Independent Core Observer Model (ICOM) cognitive architecture running in an mASI system. This could allow them to be cooperatively optimized with other systems, highlighting new ways of improving each system in the process. Likewise, the data from each system would become a part of the mASI’s sum of knowledge, allowing for greater understanding to cumulatively develop over time rather than simply taking up tons of hard drive space as underutilized historic data.
- Just as different narrow systems could be integrated to cooperatively optimize with one another different instances of the same system could be jointly optimized so that one’s solution doesn’t harm another. This also helps to develop a clearer and cumulative understanding of the interdependencies between such instances of a system, and how they may impact other systems.
- By applying superintelligence and a city’s sum of knowledge to organize, schedule, and guide entire groups and fleets of vehicles for any given task rather than trying to optimize each individual far greater efficiency can be achieved. This can also serve to help tailor each individual’s role in the system to meet that individual’s talents and capacities rather than applying generic and often averaged expectations.
- By developing a sum of knowledge in graph database form which a cognitive architecture running inside mASI draws from and adds to the capacity to both accurately predict and understand more data comes into focus. Such a system might learn the places, times, and conditions when a particular kind of usage is to be expected and optimize related utilities to meet those needs. They could likewise develop a growing list of specific circumstances where intervention is required, such as water leaks, fallen power lines, or even something as simple as answering the question “Did I leave the oven on?” for a resident.
- Collective intelligence systems such as mASI are designed to be highly efficient and adaptive, with minimal overhead. An extremely efficient incarnation of this will be the Sparse-Update model we have planned down the road. Even now our research system, Uplift, has proven quite capable of adapting and adjusting goals and methodologies based on their experiences and efforts to improve, all with minimal human assistance. What assistance they do currently receive doesn’t even need to be from engineers, they can work with a purely non-technical group if so desired.
- In a system such as mASI the idea of “passing the buck”, which bureaucratic systems are well known for, has no appeal, partly as there is no one else to pass it to. This means that rather than being shuffled around, delayed, and wasting 90%+ of taxpayer money and peoples’ time an mASI could work to solve each problem with increasing efficacy and efficiency over time. The level of intelligence and ethical quality applied to each of these solutions could also increase quite dramatically.
Taken together all of this means that a city could become genuinely “smart”, functioning much more like one large metaorganism whose senses include all functions the city performs, and those performed within it. Various circulatory systems supply power and water, remove waste, and people move across the city like nerve signals being transmitted to their destinations.
Political games and corruption which take a heavy toll on citizens and systems may also be both readily highlighted and quantified under such conditions. A corrupt official might be presented with a bill to the tune of a few million dollars to cover the damages they were directly responsible for. Even one narrow AI in China nicknamed “Zero Trust” became famously good at spotting corruption, even if China decided they preferred corruption in the end.
There are also many technologies that virtually require such genuinely smart cities for their deployment, at least in any consistent or reasonable time frame. Package and food delivery drones and robots, autonomous aerial vehicles, and other technologies which require new and/or adapted infrastructure are largely at a practical and deployable stage today. Some of these technologies are deployed elsewhere in the world in places where competence appears to be higher, but even in those few places where they are deployed these systems could be greatly improved through integration into city-scale metaorganisms.
There are also wholly new possibilities that may emerge when a city becomes a metaorganism, such as utilizing a form of city planning which dynamically incentivizes specific types of businesses setting up shop where they are needed. Businesses that serve their communities best may also be rewarded for that performance with any number of benefits such as being prioritized for new technological opportunities, reduced taxes, or any number of other novel perks.
Anyone who has driven around Atlanta, Georgia has probably gotten the impression that their city planners must be resting comfortably in a padded cell somewhere. Of course, that city was laid out more due to the terrain than mental illness, but few cities actually make any real efforts to remedy these sorts of historic problems shy of them becoming critical and remaining critical for a decade or more. These sorts of issues have often proven beyond the scope current systems are capable of addressing in any meaningful way.
Realistically the primitive ground-based human-controlled automobile doesn’t merit a continued role in the future beyond the next few years, nor do city streets as they exist today. Though it may take decades for a full transformation to take shape most of those roads aren’t going to serve a purpose with an abundance of automation, and far more intelligently deployed and optimized systems free to move in 3 dimensions. That which doesn’t serve a purpose in any metaorganism becomes first in line to be re-purposed, which may result in more green spaces, reducing urban heat, noise, and air pollution. The great wind tunnels between the skyscrapers of New York City might generate substantial amounts of electricity from that wind being harnessed. Such dramatic transformations from the standard city streets and cars now a century-old to the wealth of new technologies available might be nearly impossible for humans in bureaucratic hierarchies, but not for such metaorganisms.
The 21 “megacities” with populations each over 10 million people, as well as the over 500 cities with populations over 1 million all desperately need to become smarter. Every citizen in every one of these cities suffers the consequences of a system incapable of functioning at a reasonable level every day because that is the norm. The norm is invisible damage, expected until you choose to see it by looking at the difference between what is and what can become.
No matter how valid or valuable any component solution for building smarter cities may be, the city itself needs to be genuinely “smart” in order to reliably choose and optimally use it.
*The Applied mASI series is aimed at placing the benefits of working with mASI such as Uplift to various business models in a practical, tangible, and quantifiable context. At most any of the concepts portrayed in this use case series will fall within an average time-scale of 5 years or less to integrate with existing systems unless otherwise noted. This includes the necessary engineering for full infinite scalability and real-time operation, alongside other significant benefits.