By Logan Mullins
TAMPA, Fla. –– The World Economic Forum found that AI has the most significant impact on data-rich jobs, including software development, finance, and customer support. UPS laid off 20,000 workers in 2025 due to the introduction of new AI software. Recruit Holdings, the owner of Indeed and Glassdoor, has laid off 1,300 workers in response to the rise of AI. Google announced two rounds of layoffs due to AI.
However, many believe that AI will create more jobs than it eliminates and also boost worker productivity. The World Economic Forum predicts a net increase of 78 million jobs worldwide due to the adoption of AI. Vanguard Global Chief Economist Joe Davis believes that “4 out of 5 [jobs] will result in a mixture of innovation and automation.” Goldman Sachs later found that AI will boost worker productivity by upwards of 15%.
Employers are optimistic; workers are uneasy. But to understand the fears of today, one must look to when automation first entered the workforce.
On June 13, 1961, the U.S. Patent and Trademark Office issued U.S. Patent 2,988,237A. The patent, named for a “Programmed Article Transfer,” describes, “The automatic operation of machinery, particularly to automatically operable materials handling apparatus, and to automatic control apparatus suitable for such machinery.” Designed by George Devol and backed by Joseph Engelberger, this machine would be known as the Unimate Robot.
Photo courtesy of the U.S. Patent and Trademark Office.
General Motors unveiled the Unimate Robot that same year in the Inland Fisher Guide Plant in Ewing Township, New Jersey. The Unimate Robot unloaded finished castings from a die-cast, a particularly hazardous job in the car manufacturing plant. The robot would soon enter mass production as the Unimate 1900 Series, which would spur a global wave of automation, particularly in the car industry.
In 1966, General Motors opened the Lordstown Assembly Plant, which was one of the most efficient manufacturing plants of its time. However, in response to overseas competition, General Motors decided to redesign the factory into one that was far more automated. General Motors would spend $75 million retooling the plant. Touted as the most automated automotive plant in the world, General Motors reopened the plant in 1970. Before the overhaul, Lordstown Assembly Plant produced 440 cars per day; after the overhaul, it could produce 900.
With the implementation of the Unimate Robot, manufacturers were motivated by the increased productivity. Increased productivity means more output for the same cost. Firms adored the greater efficiency in their factories, but workers felt differently. Firms were excited; workers were terrified.
In 1964, a group of journalists and scientists wrote to President Lyndon B. Johnson about the Triple Revolution: warfare, civil rights, and automation. Workers believed that an increase in automation would lead to their jobs being taken over by machines. The slogan “You won’t get tomorrow’s jobs with yesterday’s skills” emerged. Automation anxiety became one of the nation’s top fears. Strikes became more frequent with automation being a primary cause.
However, at the time, these fears were unfounded. Automation was great for workers. From the introduction of the first Unimate Robot in 1961 to 1979, the manufacturing sector experienced consistent job growth, with over 4.4 million new manufacturing jobs created during this period in the United States. Factory workers and automated robots were utilized together to maximize the productivity of the manufacturing plants.
Data sourced from The Bureau of Labor Statistics via FRED.
The writing was on the wall. During the 1960s, the number of manufacturing jobs grew by 28%, while manufacturing output increased by over 90%. The machines were becoming more efficient than the workers.
1980 saw a transition. The Computer Numerical Control (CNC) became widespread; this device allowed a programmer to instruct a machine to do a task on repeat, with no further human operation needed. And just like that, manufacturing laborers became less efficient than machines. Their fears, once unfounded, became true. Manufacturing job growth slowed, making it the beginning of a long-lasting shift in industrial labor.
Data sourced from The Bureau of Labor Statistics via FRED.
From 1980 to 1990, automation displaced over 300,000 factory workers in the automotive sector alone. Nearly every single year, the total number of manufacturing jobs fell. At the start of the decline in the 1980s, over 19 million laborers were employed in manufacturing. By the turn of the century, the number of factory laborers had fallen to 17 million. By today, the number is barely 13 million.
On June 12, 2017, eight researchers at Google published a landmark paper on machine learning, “Attention is All You Need.” The paper introduces the concept of transformer architecture, a groundbreaking approach to processing sequences. Recurrent Neural Networks, RNNs, were the standard in processing sequences: they process data one step at a time, with each step passing information to the next step. These were great at predicting patterns and creating order, but they struggled with long sequences.
Transformers utilize a concept called “attention weights” to directly link concepts, regardless of their distance from each other. For example, in the sentence “The cat that chased the dog was black,” a transformer can connect “cat” and “was black,” whereas an RNN cannot. This simple concept of attention weighting became the foundation for large language models, the basis of modern artificial intelligence.
“Attention is All You Need” is one of the most cited papers of the 21st century, with over 170,000 citations. The core concept of Transformers has been applied in every major innovation seen today, including BERT, T5, and, most famously, ChatGPT (GPT stands for Generative Pre-trained Transformer).
Data sourced from Cornell, ChatGPT Blog, Google Blog, and Meta AI Blog. Diagram courtesy of Logan Mullins.
In 2022, OpenAI released its GPT-3.5 model, ChatGPT, trained on 175 billion parameters (roughly 117 times more than its GPT-2 model’s 1.5 billion parameters), to the public. Within five days, ChatGPT had one million users. By the start of the following year, ChatGPT had over 100 million users. Employers were quick to catch on to the power of AI tools. By the end of 2024, 78% of organizations reported using AI in some capacity. Google estimates a 10% increase in engineering productivity; JP Morgan found that coding productivity increased 20% due to AI.
We are in the early stages of AI development, but the technology evolves rapidly. Just like with automation in the 1960s, workers are afraid of AI impacting their job security. The American Psychological Association found that 38% of workers are concerned about AI replacing their jobs.
But currently, AI is benefiting workers. Gallup found adoption widespread — from finance to tech — with working AI users reporting productivity gains of up to 40%, and 77% of C-suite leaders finding productivity gains due to AI. However, the writing is on the wall.
Already in 2025, ChatGPT passed the Turing Test, an evaluation for artificial intelligence to determine human-like capabilities. As automation has done for blue-collar manufacturers since the 1980s, AI will outpace and outperform white-collar workers.
Job losses will be widespread — it is inevitable. There is simply a cold, hard truth that Americans have tried to ignore for decades. Machines can evolve fast — but it’s the people who are obsolete.
––
Photo courtesy of Wikimedia Commons.
