The world has become increasingly urbanized. The UN reports that since 2007, more than half of the world’s population lives in cities. This number is expected to reach 60% by 2030.
Growing urbanization is accompanied by an increased responsibility of cities, particularly in terms of the environment. Cities account for around 70% of global carbon emissions and more than 60% of resource use.
Simply put, the world is on a collision course with an ecological reckoning and cities are the main contributors. Therefore, it follows that they must also be key drivers of change if we are to deliver on our currently failing climate commitments. If cities are to succeed in climate action, artificial intelligence (AI) has a critical role to play.
What is AI?
AI is difficult to define, both because it covers a wide range of offerings and because it is essentially a moving target – constant learning and evolution are intrinsic to its purpose. At the most basic level, AI harnesses computers and machines to mimic the problem-solving and decision-making abilities of the human mind. Essentially, it turns human-defined goals into mathematical goals.
AI has long been touted as the technological tool with both the greatest potential for advancement and the greatest level of risk. The main risk relates to data confidentiality. Smart cities depend on data provided by citizens to function, but if that data were to be accessed by a party with sinister intentions, problems would arise. The state, too, can potentially misuse AI, harvesting and exploiting data in ways that infringe on citizens’ privacy. More blatantly, if a hacker gained access to intelligent traffic control systems, he could wreak havoc.
So how can smart cities be sure they are using AI correctly, advancing the sustainability agenda responsibly and equitably?
AI in cities
AI has the potential to impact nearly every aspect of the smart city. It enhances security through incident detection and intelligent video surveillance. It increases the efficiency of traffic management and parking on roads, as well as automated updates and tracking options in public transport. It monitors air quality, manages waste, analyzes energy consumption – and that barely scratches the surface.
To do all this, AI relies on data. Processing data, recognizing patterns, and designing solutions based on those patterns – even predicting potential future difficulties that can be mitigated – are the fundamental pillars of AI. So any city that recognizes and wants to capitalize on the potential of AI needs to ensure that its city departments collect data as efficiently as possible. This is where connected lighting can play an important role.
Sustainable partners: AI and connected lighting
Connected urban street lighting can serve as a valuable platform for a secure, distributed sensor network that can collect the necessary data that AI needs, even at city scale. Systems like Interact provide the best lighting experiences while serving as a framework to enable a multitude of smart city applications.
Streetlight sensors can monitor air quality and temperature. They can also detect sounds, such as gunshots or broken windows, then alert first responders in real time, reducing crime and helping citizens feel safer. Additionally, they can be used to streamline traffic management by offering real-time traffic information and smart parking. This information can be shared with city traffic managers or directly with drivers via an app.
Connected lighting is also essential from a sustainability point of view. If all companies and cities in the world converted all their conventional light points to connected LEDs, it would reduce annual carbon emissions by more than 553 million tonnes of CO2. This is equivalent to the amount of carbon that 25 billion trees could sequester in one year.
Smart cities that take sustainability seriously need to consider the benefits of connected lighting both as an enabler of AI capabilities and as a sustainable solution in its own right.
Potential pitfalls
AI will be essential to address social, economic and ecological challenges on a global scale. However, its limitations must also be recognized.
AI & Cities: Risks, Applications and Governance, a report published by the United Nations Human Settlements Program (UN-Habitat) in collaboration with the Mila-Québec Artificial Intelligence Institute, highlights some of these risks. “For an algorithm to reason, it must gain an understanding of its environment,” the authors write. “That understanding is provided by the data. Whatever assumptions and biases are represented in the data set will be replicated in the way the algorithm reasons and in the output it produces.
As discussed earlier, AI turns human-defined goals into mathematical ones. But if human-defined goals are based on existing preconceptions, the data will eventually reinforce those assumptions.
The AI also fails to assess its own performance. As the UN-Habitat report notes, “While it may be tempting to view algorithms as neutral ‘thinkers’, they are neither neutral nor thinkers.” The AI has no understanding of the larger context and therefore can only produce results based on its predefined optimization goals, which may conflict with larger considerations – or worse, serve a deceptive agenda.
AI systems are mathematical and cannot integrate nuances. This means that AI can sometimes end up excluding or underrepresenting subjective and qualitative information from its conclusions.
Minimizing risk through governance and accountability
There are ways to mitigate the risks associated with shortcomings in artificial intelligence. Chief among them are governance and accountability.
Accountability ensures that an entity is always held accountable – and more importantly, always feels responsible — for the impact of AI. Algorithmic systems evolve, often in unpredictable ways. A change of objective will modify their effects. Good accountability can help undo mission drift, where technologies are intentionally reallocated to surveillance and other extraneous purposes. It can also help ensure that bad faith actors are not able to deliberately mismanage AI goals or redirect them over time.
AI governance refers to the sum of AI regulations, ethics, standards, administrative procedures, and social processes. Governance helps ensure that AI is used in an inclusive and equitable way, and that preconceptions or lack of awareness in the early stages do not allow AI outcomes to widen the digital divide or exacerbate existing inequalities. Governance allows local authorities to assess the opportunities and risks offered by AI, so that they can then apply it according to the local context.
Consultation with citizens and communities is also vital. The public is the main actor in every city; they need to have a say in how a tool as powerful as AI is used in their community. This helps ensure that the AI solves local problems, without making them worse.
Responsible AI
The ability of AI to generate and expand the possibilities of smart cities is considerable, especially to advance sustainable causes. There are risks, but also ways to circumvent them. Conscientious decision-making that considers local communities and consults with local authorities will help ensure cities get the most out of AI.
To learn more about Interact, click here.
#Enable #smart #cities #connected #lightning