Artifical Intelligence and Labor

I. The Issue

One of the biggest inventions of this century has been Generative Artifical Intelligence. It is still a fairly new technology, but many great leaps have been made in the past few years alone. As an example of how much AI has developed recently, the Will Smith Eating Spaghetti videos serve this purpose well. The reason for such remarkable progress is largely due to the massive capital inflows towards AI, with roughly 1.87 trillion dollars invested in AI since 2020 according to Stanford's 2026 AI Index Report. This sort of investment does not come for every new technology or innovation. In order for there to be this much investment, to where articles calling it a bubble are being written every week, there must be a heavy financial incentive. While there does remain the speculative angle, investors trying to cash in on the hype for AI, there is a legitimate business angle as well.

When people think of AI, they often think of ChatGPT, Gemini, DeepSeek, etc., conversational AI. While these are very front-facing, they will likely not be the primary focus for AI. People love asking AI questions, with Pew Research Center reporting that one in every four Americans use AI on a daily basis. However, as Stanford notes, most models operate on a free or nearly-free basis; if these models charged money, people would use them far less. The main source of revenue for AI companies will not be direct-to-consumer models, though some attempts to keep these available may persist by larger firms such as Microsoft and Google, but rather business-to-business, enterprise, products.

Stanford again reported that 88 percent of businesses use AI in some business capacity, while about 35% of businesses with over $1bn in revenues expect a significant decrease in their employee count. The site Layoffs.fyi automatically tracks already, in the tech sector, significant layoffs. Below are layoffs as reported by them, as well as total computer employment data from the BLS's annual May OEWS tables, the combined employment data among four big Silicon Valley firms who are heavily involved in AI as per Macrotrends.net (Google, Meta, Microsoft, and Oracle), and the amount of AI investment as reported by Stanford:

A few patterns emerge from this data. For one, the problem of layoffs in tech seems to be more correlated with general investment in AI. The only periods wherein employment in the computing sector and the AI-focused Silicon Valley firms did decline overall was 2023, where the overall investment into AI slowed down as compared to the previous year. So while layoffs do occur, it seems to be less related to employees directly being replaced by AI, and more related to those employees only having a job due to AI related investment. In other words, these jobs that were lost only existed due to the AI investment in the first place. Moreover, the surge between 2021-22, when AI investment decreased, suggests that perhaps these firms may have overhired in the first place. However, there are ultimately more factors than just AI determing whether companies hire more computer employees. Still, it is a significant factor.

Many of the places AI does indeed replace people are not in the back-end of the tech sector itself. These are more normal jobs. For example, Amazon Delivery Drones currently operate in ten American cities, with five more to come. Starship operates in 300 cities and has completed over 10 million deliveries, while Servero operates a fleet of 2,000 robots, delivering food via apps like UberEats. Uber itself indirectly operates self-driving taxis in four cities, Waymo, a Google company, directly hosts thousands of self-driving taxis in fifteen cities, and Tesla's Robotaxi operates in seven cities. In essence, much of the unskilled gig-work jobs are liable to be replaced, especially in highly urban environments, some of the only places where these sorts of jobs are viable as standard income to begin with. Deloitte reports over 57 million Americans engaged in some way or another with gig work (Amazon Flex, Uber, DoorDash, etc.). This is a highly volatile sector of the economy, with few benefits and poor pay due to workers only being technically contractors. These are often workers who do not earn enough at other jobs (or possibly cannot find an other job) and instead must make ends meet via this sort of work. Already somewhat excluded from the traditional production process, the encroachment of AI in these markets could drive these people into even more dire financial situations.

Other than gig work, there are many sectors where AI could be highly impactful. Call centers, for one, employ 2.8 million workers in America. Already, there are services offering AI-powered call center operations such as NextLevel, and AI-powered virtual assistants have become commonplace on a broad number of websites for small businesses as well as large corporations such as Amazon. These sorts of AI would vastly cut down on the required staff for these companies. These employees would then be pushed to lower-paying jobs, or perhaps gig work -- already a crowded space.

In the 1960s-1990s, there was the period of deindustrialization in America, where many manufacturing jobs were shifted overseas to nations such as China, Vietnam, and Taiwan. The reasoning for this was obvious even to bourgeois economists: they worked for less. The American workers were frequently working in factories for decent wages, enough for a single income to support a full family. In nations with a far lower cost of living, such as China, workers did not need as high of a wage. Once profitable investment opportunities were dried up in America, trade barriers such as Bretton-Woods were broken down and capital outflows began. This left many Americans totally destitute, with many towns in the industrial northeast turning into a shadow of their former self. For example, the now-defunct Bethlehem Steel Works employed 31,000 workers. The Wisconsin town of Janesville held a General Motors plant, whose closure is agreed to be a primary cause of the decline of the town. Not only are workers directly affected, but subsidiary businesses. For example, the Janesville plant's closure also affected diners where workers would eat, shops where they would purchase luxury goods, leathermakers and steel mills making the material for the automobile, and so on. Nowhere is this effect more pronounced than in West Virginia. West Virginia was home to 130,457 miners in the coal industry. Nowadays, this number is closer to 20,000.

• Marketing Analytics (

II. Theory and History

III. Consequences