Introduction
There are many arguments to be made against the use of AI, especially in the workforce and educational system. In this post, I argue that AI usage could be damaging to the foundations of our current institutions. I must note that I prescribe myself to the belief that AI is a positive and helpful tool that can be implemented in our daily lives. For the sake of this assignment, I will offer common arguments against LLMs. Please also note that this post was made in collaboration with Sean Gogolin, who took a stance in support of AI usage.
Replacement of Human Workers
One of the most common arguments I hear against AI is the concern that LLMs and other forms of augmented AI will replace human workers in various industries which will lead to economic inequality and skill disparities. This concern has been especially relevant for individuals in the trucking, manufacturing, and data entry industries where large amounts of work are completed on computers or are at risk of being replaced by AI.
One of the major problems that arise when workers are ousted from their positions is a lack of skill or knowledge to transition to new roles. To combat this issue, some companies have invested in retraining programs to help workers who are displaced get reacquainted with a new position. The problem here is that the success of retraining programs has been limited. In other words, some companies are finding that investment in retraining is not worth the time or expense. Instead, I argue that companies can simply hire new individuals for the job and pay off the old employees.
Further, CNN finds that AI is replacing jobs faster than people think. These jobs are ones that are more niche, creative tasks, invoices for reporting, crafting job posts, and writing press releases. All of this reliance on AI is to reduce costs, CNN notes. In this view, AI poses a significant threat to the human task force, which will only be further replaced as company’s lean LLMs to automate tasks that were previously carried out by skilled workers. Some might argue that AI only stands to benefit individuals and assist with current workflow. This being said, one must wonder what happens to those workers such as truckers or others that lack skills to go into other industries. Are these workers to be paid off? Are they let go with severance packages? What if a worker just started and loses their job to AI?
AI Makes Mistakes
Along with the replacement of workers, I also argue that AI is very new, and thus, very unreliable. In January of this year, a Minnesota Attorney General named Keith Ellison was found to have included AI-generated citations in a court filing. In December of last year, BBC complained to Apple after Apple Intelligence generated a fake summary of one of BBC’s news stories. This story was written about the sensitive case of Luigi Mangione and the killing of UnitedHealthcare’s CEO Brian Thompson.
The reality is, AI seems to infer a lot of its information based on past happenings and a false foundation of ‘common sense’ reasoning. Most AI models are also found to be too quick to please and possessing of out of date information. This means that people and businesses are relying on faulty software to complete important tasks.
A writer from the International Director named David Winter wrote in his article that as machines become more advanced, their errors can become more serious. Winter goes on to state that the algorithms that AI uses can be prejudiced. He offers the example of an AI system designed to predict the likelihood of an offender committing a crime in the future based on race.
AI use can be damaging in other forms as well. For instance, a Tesla car smashed into the back of a tractor-trailer, killing the driver after a mistake with the auto-driving feature. The problem here is twofold. One, AI has no sense of morality or ethics, and is instead, only driven by algorithms, patterns, and code. Two, AI relies on human creation so much so that its data carries the biases and irrationality of humans. This means that AI may be faulty in the way that it sees patterns where there are not any and does not know when an answer is wrong. Even scarier perhaps in the idea that AI may refuse to admit that an answer is incorrect while humans blindly follow.
Overarchingly, AI has significant negative implications for the current workforce and broader society. Humans have started to lean on AI for more and more important tasks, yet, not enough people have stopped to wonder if AI is reliable. Experts have already found flaws in the reasoning and output of AI models. As technology gets more advanced, these flaws may become more readily apparent.