What AI Can Tell Us About Happiness

February 28, 2025

If you want to know how well a society is doing and whether its citizens are happy, you might look at things like life expectancy, literacy rates or income levels.

These are all examples of what researchers call quality-of-life measures. Politicians and governments have used them for centuries to assess the well-being of their citizens.

Now, researchers are applying artificial intelligence to the study of well-being, searching for deeper answers to what makes countries, cities and people happy. A new study in IEEE Access investigates how machine learning and large language models (LLMs) analyze research on quality of life.

More Than the Economy

While quality-of-life indicators don’t directly measure happiness, they are often used as a proxy. They help policymakers, researchers and organizations understand social, economic and environmental conditions, guiding decisions that improve overall well-being.

Traditionally, they have focused on a single number, usually related to the economy. Gross domestic product, for example, is frequently used. It measures the total economic output of a country, and it has served as an indicator of quality-of-life based on the idea that if a country is wealthy, its citizens likely have a better quality of life.

Over time, however, researchers realized that relying on a single metric to evaluate society’s well-being wasn’t always effective because money doesn’t necessarily buy happiness.

According to IEEE Member Ning Hu, happiness is often comparative rather than absolute.

“Happiness is fundamentally rooted in our sense of self-identity and understanding of our position within society,” Hu said.

To get a fuller picture, researchers began combining several metrics into one understandable number. For example, the U.N. Human Development Index looks at life expectancy, education levels and the average amount of income each person in a country receives, translating all of those into one number. Other indices use other metrics, all with a goal of helping researchers and policymakers make easy comparisons between countries.

And quality of life measures aren’t just used to assess the well-being of nations. In medicine, they have been developed to evaluate the health of individual patients and make predictions about their future health.

Some quality-of-life indexes are designed to assess cities on dimensions like quality of government, wastewater treatment, recreational opportunities, and economic outlook.

The Rise of AI

As computers became capable of storing and analyzing ever-larger datasets, researchers began applying big data techniques to the study of human well-being.

One bright spot for researchers has been the rise of sentiment analysis, which examines communications like social media posts for opinions and emotions to determine whether they express positive, negative or neutral attitudes.

That work has often come to interesting conclusions. For example, when negative feelings increased on one social media site during any single day, the price of gold tended to increase. In the health field, sentiment analysis has also been applied to doctor’s notes to be included in quality-of-life indicators.

Researchers in Korea used machine learning to predict whether countries were happy or not. They found that several highly developed countries weren’t very happy, while several less developed ones were. To increase happiness, the authors found that policy makers should lower youth unemployment rates and improve healthcare quality.

Another analysis found strong relationships between national happiness, high average IQ and high concentration of people who scored high in the sensing and judging fields of the Meyers-Briggs personality test.

Can AI Make Us Happy?

Given the amount of research into quality-of-life and happiness, it’s reasonable to wonder whether AI can make us happy.

The authors of the IEEE Access paper examined 87 pieces of research that applied artificial intelligence to the study of well-being. They found that while there were some effective methodologies, they found several research gaps and areas for further research.

And while AI can point decision-makers in the right direction, it cannot, ultimately, lead any one individual to happiness.

The ideal role for AI is as an enlightened tool for expanding our understanding of well-being,” said IEEE Senior Member Eleanor Watson, “leaving the ultimate pursuit and definition of happiness to humans themselves.”