KArtificial Intelligence (AI) has the potential to turn our lives upside down. And it can also answer questions that have been inaccessible to us until now. In economics, for example, researchers use them to explore how immigrants become economic climbers or how climate change and economic development are transforming remote regions of the world. Susan Athe, a pioneer in the use of the new technology, wrote back in 2019 that machine learning – the core of AI – will have a “dramatic impact” on economics within a short period of time. The conclusions he and his colleagues draw based on the new methods, in turn, change our understanding of the world.
Machine learning can help in both compiling and analyzing data. In the former case, it replaces work done first-hand by a large team of research assistants. And in a fraction of the time. This makes very large data sets usable, for example, by economic historian James Feigenbaum. Boston University Knows how to use it in your research. It uses historical census data to track people’s careers over time. The data contains a wealth of information about education, family relationships and economic status. By linking different data sets from one census to another, Feigenbaum and his colleagues were able to show, among other things, how immigrants integrated into the United States and achieved economic success.
The difficulty is that this data is usually fraught with errors. It is not so easy to track the same person from one decade to another. German immigrant Schmidt became Smith within ten years. Or a transposed digit creeps in with the date of birth. To create a coherent data set without the help of computers, tedious manual work and a large number of visceral decisions would be necessary. Feigenbaum trained an algorithm by adding people by hand to a small sample and thus teaching the software what common alternative spellings are, for example, and what tolerance for deviation is still acceptable. This algorithm could then be fed with data from millions of people and found the correct pairs with a relatively high degree of accuracy.
insights from remote areas
Sometimes, however, economists have to start a step earlier in empirical research. In the past, if the data was not laid out neatly in a computer spreadsheet, but was simply printed on paper, one had to type it painstakingly by hand. This is not only very expensive, but error-prone. Economist Melissa Dale has developed a program that, with the help of so-called deep learning, can not only recognize numbers in scanned tables, but can also accurately digitize their layout.
Freelance twitter maven. Infuriatingly humble coffee aficionado. Amateur gamer. Typical beer fan. Avid music scholar. Alcohol nerd.