Artificial Intelligence: How far is too far?

While there is a natural fear of futuristic technology rendering human workers obsolete, there are limitations that will keep AI in check for years to come

Artificial Intelligence, or AI, is an area of technological advancement that’s gaining interest in various sectors to make daily life more convenient and efficient. While AI may come with the typical “sci-fi” fantasies portrayed by the media, in reality, it is already sneakily at play in different aspects of our day to day lives. The numbers speak for themselves, as according to a recent study, the number of businesses using AI grew by 270% in the last four years. From online dating to the way we shop, it has come to substantially dominate how modern businesses are run. As with most major technological developments throughout history, it is natural that the increasing use of AI comes with the fear of rendering human workers obsolete. 

Before diving into the implications of global AI usage, it is important to understand what artificial intelligence is and what it is not. The Britannica definition of AI is “The ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings,” which is a rather scary interpretation. AI is used to describe computing, systems, and technological advancements that can perform tasks associated with human reasoning. It is more than a regular algorithm, which is defined as a simple “formula” or set of instructions for the processing of data. AI takes this processing to the next level, as the technology itself has the ability to learn and the capacity to modify, adapt, and change in response to data input. This is the “intelligence” aspect of the technology that makes it much more advanced.

“The natural worry that quickly becomes eminent is that if this trend continues spreading at an exponential rate, will jobs that we once took for granted quickly disappear?”

An example of AI being put to work on a large-scale was the opening of the “Amazon-Go Grocery” store in Seattle, Washington in February of 2020. The massive 10,400 square foot supermarket is more reliant on technology than conventional shops to operate the facility. Rather than talking to the local till worker or shelf stocker, customers scan a code at the entrance and proceed with shopping and paying straight from their device. The natural worry that quickly becomes eminent is that if this trend continues spreading at an exponential rate, will jobs that we once took for granted quickly disappear?

In the case of Amazon-Go, a spokesperson for Jeff Bezoz, Chief Executive Officer of the company, assured that this would not be the case. However, on the opposite side of the spectrum, Union President Marc Perrone quickly contrasted Bezos’ sentiment commenting, “Bezos wants to create stores that serve food and groceries and eliminate the jobs real people need.” Current trends in AI indicate that this fear, while natural, is not anything to worry about for a very long time. A recent paper published in Massachusetts Institute of Technology’s Task Force titled “Artificial Intelligence And The Future of Work” gives a much more optimistic view of AI and its future implications.

This paper predicts that rather than rendering human labour obsolete, AI will continue to drive innovation in many industries and create opportunities for new areas of growth, thus creating new jobs. In addition, there are limitations to AI that help ease the fear of it being able to replace regular workers. AI programs are typically only capable of “specialised” intelligence, meaning that they are usually limited to solving one problem at a time. This creates rigidity and prevents “learning” outside of the system’s programming – an area where we as humans can beat AI.

“Surprisingly, AI is being used incredibly often in the hiring process.”

There are other factors that can limit the ability of AI to take over too quickly. The “learning” aspect of the program often requires mass amounts of highly specific data, which gives rise to issues of privacy and security around the information needed to construct and run these AI programs. They are also costly to run, as the necessary electricity to run a one language AI model is estimated to be nearly €4 m. It is still more profitable for current business models of many companies to use workers on a daily basis over this advanced technology. 

Surprisingly, AI is being used incredibly often in the hiring process. Rather than eliminating jobs from people, AI programs are used to sort through applicants and assist hiring managers in their tasks. In fact, up to 75% of resumes are rejected by an automated tracking system (ATS) before they even reach a human being. Rather than pouring over LinkedIn profiles for hours, 67% of hiring managers in 2018 reported that the use of these ATS programs were making their jobs much easier. AI can be employed to assist in people’s jobs that they already hold, rather than replace them entirely. 

Similar to the concern of AI replacing workers, is the capacity of privacy breaches with these programs. The European Union has regulations in place in order to ensure that businesses who choose to use AI technology are adhering to strict guidelines of what is acceptable versus unacceptable. These laws and regulations are another added layer of protection from AI advancing rapidly and overwhelming society and businesses. The EU analyses and separates AI systems into three categories based on their risk – unacceptable, high, and low-minimal. Some examples of high-risk systems would be those that use biometric identification and low-minimal risk systems would be the familiar AI chatbots on many websites and social media pages.

While AI has become increasingly popular in business models and our day to day lives, the fear of it stealing all jobs and essentially replacing people is nothing to worry about for now. The limitations that keep AI from advancing too quickly are enough to keep this technology in check. Actually, according to an Oberlo study this year, 62% of consumers were willing to submit data to AI in order to have better experiences with businesses. This shows that there is some trust with these systems, as there should be.

Shannon McGreevy

Shannon McGreevy is the Online Editor of Trinity News and a Senior Sophister student of Biochemistry.