This article is a part of the Holos Project, a four-way partnership (ENGin, The Los Angeles Times Insider, Published Points of View, The Outspoken) established to empower the voices of Ukrainian students across the world through one-on-one journalism mentorship. This article was written by Zlata Kononova from Ukraine with the mentorship of Ryan Geng from the United States.
“Holos” is the Ukrainian word for “Voice”.
Artificial intelligence is a sphere of enormous opportunities and great risks. Much of it is not yet known and not fully understood by the global community, which has caused a wave of mistrust and fear. This fact is also made more terrible by films that depict the end of the existence of mankind provoked by the aggressive actions of AI, like “Terminator” or “Next”. But at the moment, artificial intelligence has been demonstrating the ability of a computer to imitate and reproduce human intelligence and its behavior. At this stage, it will not be able to surpass people or replace them, because there is not yet AI that will create a lot of similar ones on its own. That means that for now it completely relies only on our knowledge, developments, and discoveries. Generally speaking, AI simply cannot replace us, because it is not able to entirely independently deal with any processes, relies on a database created by humans, and is devoid of emotions.
Firstly, a neural network cannot surpass humans in art and creating something new. During the creation of drawings, sketches, and other creative products, AI relies on a database that has already been created by people and cannot come up with something radically new or make a breakthrough in something. AI can only work in a couple of ways. The first way is it analyzes its database, which has already been created by someone, and finds matches with requests. The second is completing a task following a certain set of human-written rules, such as a language translation or a Google image search. So, at this stage, I see it only as a source of inspiration for creative people. With the help of a neural network, everyone can look at their work differently and be even more inspired by noticing details that they have not seen before. Also, at this stage, neural networks can only do sketchy work. Looking at all pieces of art created by artificial intelligence, it is obvious to see a lot of defects. Certainly, it is possible to get a more realistic result specifying more suggestive tags so that AI recognizes them and does everything according to the orders and instructions. However, even under such conditions it will not be able to do everything exactly the way you imagined and wanted to get. Therefore, artificial intelligence and neural networks can do only sketchy work, and professional specialists who are well-versed in this topic will always be needed because only they can create exactly what will fully correspond to the description of the customer.
Secondly, at the moment, AI cannot entirely independently deal with any processes as it only automates them, helping people avoid monotonous work. For example, let us take hiring staff and those things that companies spend too much time on, such as sorting resumes, scheduling meetings and interviews, making phone calls, and sending emails. AI can handle these repetitive tasks with ease, and pick out the best candidates, who have the highest chance of success in a particular organization. But, ultimately, it deals only with repetitive work without making something new or replacing people from this work area. People are still needed as employees. Release of workers from monotonous tasks will allow them to devote more time to more important tasks; for instance, developing relationships with the best candidates. It helps free up people’s time and energy in order to focus on more important activities while improving their quality of life. Precisely, healthcare is one of the directions in which AI has made great strides in helping people. AI algorithms can be trained to analyze medical data and images so that doctors are able to make the most accurate diagnoses quickly. Neural networks are also used to improve transport systems, making them safer and more efficient, starting from controlling traffic lights and trains to the autopilot of today’s cars. AI also helps creators improve video games by adding small details to their environment and improving the quality itself. Again, AI is just doing repetitive work and tasks without creating anything new or making breakthroughs. According to the “Accellabs” website, such areas as retail, and financial services will extract the greatest benefit from AI. It is about increasing productivity and improving quality and consumption. Also, it will affect economics in the main topics currently being discussed. For instance, productivity increases through the automation of business processes (including the use of robots and self-managed transport), the strengthening of existing working resources with the help of AI (intellect, which helps and expands the capabilities of the human brain), and increased demand due to the availability of personalized or equipped AI goods and services.
The next reason is the number of mistakes that AI makes due to its limitations and issues. First, they are devoid of feelings. An AI system designed to maximize a specific goal like profit and employee efficiency can make decisions that will directly harm people. There are already specific examples to prove this statement. One of them is Amazon, where the work of warehouse employees has been fully regulated by a neural network. This caused a lot of scandals with employees who had to do the impossible because the neural network set tasks that were unattainable for a living person. As was said on the website “The Guarder”, “‘I was a picker and we were expected to always pick 400 units within the hour in seven seconds of each item we picked,’ said Espinoza. ‘I couldn’t handle it. I’m a human being, not a robot.’” Ilya Geller, who worked as a stower, told of the pressure workers face from being surveilled by computers to ensure productivity rates are met. “‘You’re being tracked by a computer the entire time you’re there. You don’t get reported or written up by managers. You get written up by an algorithm,’ said Geller. “You’re keenly aware there is an algorithm keeping track of you, making sure you keep going as fast as you can because if there is too much time lapsed between items, the computer will know this, will write you up, and you will get fired.’” Moreover, they will act badly in the service sector. Many companies use AI, but they are not able to make exceptions that human employees can make based on a live conversation: seeing the face, facial expressions, and emotions of the client. Second, the system is not responsible. Who is responsible for the mistakes made by AI? They can make decisions without human intervention and no one can hold them accountable. It is programmed to independently make unique decisions depending on various factors and it will not be possible to punish its creator. The next issue is AI bias. Its system is only as good as the data it takes information from. If the data is biased or true only by a small percentage, then the system may also be the same. According to the YouTube video made by “Itpedia”, taking ChatGPT as an example, the main advantage is that it remembers the previous results of your conversation and can answer any question using a database from the Internet, which makes many people think that this program can replace Google, but it is not like that. The main problem is that when you enter a question into Google, there are a large number of resources and various answers, which cannot be said about ChatGPT which gives only one specific answer. At the same time, if you enter the same question into the search line, AI can answer differently. For example, in the case of a historical question, it will not be able to compile from several articles and will select only one. How in this case will AI determine which of them is correct? For certain applications, this can be a big problem and can lead to unfair consequences.
All in all, AI has no possibility to completely replace humans due to the objective reasons said above, specifically, lack of emotions, inability to operate independently, relying on work initially created by people.