The benefits of Urban AI

AI has enticing potential - but we should not try to run too fast...

AI has the potential to support a wide range of activities and services, from supporting the teams that oversee cities and towns through to helping the people that live and work in them. It can potentially help with all kinds of urban matters - such as decision-making on new directions to take, safety and security management, transport, energy and water management, citizen support and business engagement, to name just a few areas.

Just like adopting an approach to masterplanning that involves everyone is the best way to start or refresh an urban plan, city and municipal teams should think about how they can involve everyone in the early discussions on how they could potentially use AI, before a plan is put together. Experts and AI Thinkers should be involved in these discussions. We should capture the views and ideas from everyone, from citizens to corporations, AI start-ups, civil society organisations and academia.

Any approach to using AI needs appropriate investment and funding, and a good investment case should be made which ties together near-term action to deliver quick results and long-term plans. 

What are some of the possibilities...?

While generative AI chatbot assistants are currently among the typical AI solutions being used by cities, some observers believe that other aspects of AI solutions including "large language models" could provide holistic support platforms in which urban activity can be tracked and analysed, to support decision-making on an ongoing basis on how to ensure the urban place is optimised to be the best that it can be - with value always being the driver. If value can be achieved through good governance and data management and protection, it could provide a comprehensive view of a city or a town, allowing municipal teams and others to see valuable and value-driven insights through visualisation of the urban system. 

Cities and towns are made by people. If data can be captured in a data-protected way and used to improve decision-making for planning to make places better, AI and data analytics can truly help people to live better lives, and businesses in urban places to prosper in. a meaningful way. This data can be much broader than incidents tracking and events correlation, air pollution monitoring, and traffic monitoring and analysis. It can potentially be all sorts of "open data" combined with data that is specifically captured using certain devices or from discussion groups.

A very brief overview of AI

General information about AI is available from the ISO.

AI solutions work by processing large amounts of data to identify patterns, provide consolidated analysis, and provide support for decisions. The process often involves using algorithms, which are rules or instructions that guide how an AI solution works. Good data to "teach" an AI algorithm is key. Machine Learning, which is the ability of a computer system to learn from data without being explicitly programmed, is already being used by many organisations to improve how they operate. 

For an urban environment, there is great potential for AI to support the creation of data maps and linkages between data that have previously been too hard, too time-consuming and too costly to produce.

Aiming to stitch AI into daily activities, to make them better...

An urban team should always be learning and responding to the needs and changing habits of the people who live and work in their urban areas. For example, the processes for how they receive and respond to service requests and suggestions from citizens and businesses, and how they receive citizen and business feedback to act upon, has the potential to be improved by the appropriate use of AI.

AI can support urban matters ranging from "structured data" analysis such as air quality monitoring, traffic patterns and energy management through to "unstructured data" analysis such as talking with individuals and groups about how to improve urban areas, and structuring conversational text into a way that can provide outputs to support decision-making and meaningful action much faster than it would take people to do. It could potentially be valuable for monitoring key urban indicators that require continuous monitoring, with alerts activated when thresholds are reached.

A few challenges to be mindful of...

Challenges we need to consider before using AI for urban development include data privacy concerns, implementation costs, how contracts with AI providers can best be set up with gain and pain sharing between cities / localities and the AI providers, and the vital need for good AI governance and ethical rules in AI use.

Some considerations for cities and towns include:

1. Citizens - how can it help them in meaningful ways, in a people-first approach?
2. Businesses - how can it help them in meaningful ways, with efficiency and effectiveness?
3. The city / municipality team - what could it help you with, and what would be the highest priorities to focus on first? Can it help you diagnose, track and monitor what is going on?
4. The city / municipality team - always use systems thinking when considering the use of AI, and involve everyone in early discussions. Do not create a plan, then share it for "consultation". 
5. The city / municipality team - if you are to engage with AI specialists, make sure there is mutuality in agreements and contracts, with a shared accountability to deliver value.

A few points to think about when considering the potential for AI use in a city or a town: 

1. Good governance needs to be in place to ensure accurate data collection and good quality analysis, and also data protection and security.
2. We should take a systems approach to having a plan to implement and use AI. This may involve taking an approach such as "Think Big, Start Small and Scale). AI experts may be able to advise on this.
3. Involve everyone in the discussion early, not once a plan has been drawn up.
4. A plan to use AI needs appropriate investment and funding, and therefore a proper investment case needs to be made. A full understanding of costs including licensing models is key, otherwise you will pay more than you think.
5. AI solutions typically require a lot of computing power, which is energy intensive. Think about this as you develop a plan.
6. As we describe in the website section on using urban data, whilst we want to collect good data, we must avoid the "data paralysis" trap, in which we continually seek ever more data before taking action.
7. The judicious use of AI can add value - but this does not mean that it should be applied to everything, in a belief that it can "fix everything" (it will not).
8. The use of AI must therefore be based on meaningful targets and meaningful benefits that deliver value.
9. A good way to either make a start or refresh what you are doing with AI is to see what other cities and towns are doing, and learn from them, and also to learn from AI start-ups and businesses.

As part of involving everyone, ideas for using AI could be displayed in Urban Future Centres, for everyone to see and to ask questions about. Workshops could be held. This approach as part of urban governance and adhering to an agreed Charter for how the city / town works could be a valuable way of achieving high performance outcomes, demonstrating transparent and open governance, and ensuring accountability for actions.

A few city examples

Cities around the world are pursuing AI initiatives. Below are just a few examples (the author of this website has no affiliation with any of them).

Buenos Aires:

• Argentina’s capital city launched a chatbot, "Boti", in 2019. It has evolved with generative AI,. In January 2022 it achieved 11 million conversations to become “a preferred channel for citizens”, according to the Observatory of Public Sector Innovation.

Singapore:

• The Lion City state was the world’s first nation to create a digital twin as a virtual model of its landscape. Following a government initiative launched in 2023, it has come up with more than 100 generative AI solutions.

Boston:

• Boston in the US would like to become more bike-friendly. Planners have been looking at learning from cities in Europe such as Copenhagen, but locals are concerned about how such changes could impact their city. Gen AI is helping Bostonians envisage a different layout for their city with a greater focus on cycling as part of active mobility.

Dallas:

• Autonomous vehicles have already been trialled with varying degrees of success. A new breed of self-driving lorries is being tested between Dallas and Houston in the United States state of Texas which are trained on generative AI.

The SDG AI Lab

The SDG AI Lab is a joint initiative of UNDP Nature, Climate, and Energy Team, UNDP Finance Sector Hub, and UNDP Istanbul International Center for Private Sector in Development (IICPSD).

Urban AI

Urban AI is a Paris-based think tank that focuses on the emerging field of “Urban Artificial Intelligence.”

Other AI services

Forbes provides an "AI 50" list.

Many AI start-ups around the world are doing interesting work.

You probably have to prioritise what to address first, and what needs to come later, in an AI Plan, which should be linked to your overall urban Vision and a Data Management Plan. 

Think about involving everyone in your city or town about how to get started.

Then agree how to start in a small way and  how to scale up your data (in a data-protected way) over time. Have a flexible plan which can adapt as circumstances change.

Consider how to create a compelling Investment Case to obtain funding for your AI plan. With careful thought, there could be good opportunities to raise capital and ensure resources for ongoing maintenance, if you can demonstrate value in the Investment Case.

Consider guidance from AI frameworks and standards, which are evolving.