What we know two and a half centuries later is that AI can play a critical role in fostering broad-based economic development — and not just in terms of creating new industries, although this is no small thing. Generative AI alone could add $2.6 trillion to $4.4 trillion to the global economy. The potential of AI, in all its rapidly developing varieties, is even greater.
One way to prosper is to identify areas where a country, region or city has a competitive advantage. AI can perform real-time analyses to spot emerging opportunities. For example, an East Asian city used analytics models to identify two high-potential areas: beverages and auto parts. It focused its development efforts in these two sectors and saw per-capita GDP grow by $8,500 in just six years.
Or consider how natural disasters can damage economic development. The use of AI-driven spatial data and satellite imagery can provide critical warnings of impending events, saving lives, property and hope.
Next, think about what it takes to develop the most valuable resource of all: human capital. AI can be used to identify changes in the labor market, such as which jobs are likely to be most affected by automation and where future ones will be clustered. Then it can be used to provide the training that workers will need to succeed.
Finally, AI can provide real-time measures of economic activity, enabling leaders to spot small problems before they become big ones.
In these ways and others, AI can make economic development smarter. But it does not confer wisdom or judgment. Countries, regions or cities that are rich in data will still need to sort through the noise to make good decisions.
In addition, AI-enabled insights are just thought balloons if there is no follow-through or if conditions on the ground do not support the indicated action. Cincinnati used predictive analytics to crunch a wide variety of data points to spot possible investors. The AI-enhanced effort that provided these insights, known as REDI Cincinnati, is credited with helping to attract some $6.2 billion in capital investment over the last decade. It was local leaders who closed the deals, for which they needed to show their region had the talent and infrastructure to meet investor needs.
To turn AI’s development potential into economic growth, a few principles will apply widely:
Establish trust. To build confidence and trust among governments’ employees, it can be helpful to find a few quick wins. And it’s important to make the case that AI is meant to enhance expertise, not replace the experts.
The public needs to be won over, too. A survey conducted in 31 countries last year found widespread nervousness about AI, particularly in Europe and North America; only a third believe its use will improve their lives or the economy, and about the same think it could replace their jobs. Half worry about their personal data.
For both constituencies, familiarity can breed acceptance. It is important for governments to make the effort to explain AI. One model is Britain’s Algorithmic Transparency Recording Standard, which sets out how its public sector’s algorithmic tools are created and used.
Don’t be afraid to use partners. AI is changing constantly, and few governments have all the AI talent they need. At the simplest level, it may be possible to buy macroeconomic models and other tools off the shelf, rather than starting from scratch. It can also be helpful to work with other government agencies or external providers to build expertise and fill gaps. One model might be the program the U.S. government has for IT and cybersecurity staff to rotate to other departments for six months to a year: They learn on the job, and also bring their own expertise, before going back to their original post. The result is that different departments learn from each other.
Make data seamless and accessible. Bureaucracy, politics and regulation can make it difficult for data to be put to work, hobbling AI’s potential. In the European Union, a 2022 study found that 6 out of 10 public-sector AI use projects hadn’t gotten beyond the development or pilot stage. And the data itself needs to be accurate and usable, something that is not as common as it should be.
Data governance needs to be a high priority, so that government data sets are available, readily shared and secure. In addition, it can make sense to acquire outside data for specific purposes. The World Bank, for example, used mobile phone logs to identify the poorest households in Afghanistan; this was just as accurate and much less expensive than field visits. A side benefit is that partnering with local organizations this way can also build trust.
Data is not everything, and machine learning is no replacement for human judgment. But decisions informed by data can be better ones. So it makes sense to deploy AI and other advanced technologies for economic development. Adam Smith would probably agree.
Tarek Mansour is a senior partner in McKinsey & Co.’s Dubai office, where Andreea Zugravu is a partner. Ben Safran is a partner in Washington, D.C.
Governing’s opinion columns reflect the views of their authors and not necessarily those of Governing’s editors or management.
Related Articles
-
Future in Context
