Before I delve into the details of this article, I must warn that this piece is likely to receive a lot of criticism, as it contradicts the generally held views on AI and AI agents within the tech industry and the public. Like the internet revolution that happened probably 20 years ago, the market is witnessing the boom of artificial intelligence. However, unlike the internet revolution, I believe AI receives too much hatred.
Why is that the case when it is merely just a tool that is put in the world to aid us?
With the development of AI, many people complain about losing their jobs or are scared as to whether they will ever get replaced by an AI agent. However, that is not the likely scenario, as AI is merely a tool that exists on the internet, and if used correctly, it could help boost productivity to a higher scale than imagined.
But how do I account for the hallucinations and biases in the answers that an AI model provides?
Honestly, I do not hold any accountability, but since I am writing this article, I believe all I can share are opinions about a subject. ChatGPT models are often reported to generate biased and hallucinated answers, which stands as its major drawback. It works on a learning model known as reinforcement learning, which works like a reward/penalty-based system. When the user is satisfied with a response, the model gets a reward, and if they are not, the model receives a penalty. This learning mechanism gives models such as ChatGPT to learn from experience and train itself based on that experience to improve its performance to give the utmost user satisfaction. A point that we miss to see is, as human beings who have been socially and intellectually evolving for at least 1000 years, we still do not have the answers to everything in the world. We are taught answers, or we end up seeking answers, which is also a result of multiple people’s opinions. However, people are biased, and these people are the ones feeding data to AI. How could an AI’s opinions or answers be unbiased if people’s opinions are biased by themselves?
But why else are people losing their jobs?
From an employer’s perspective, people have lost their jobs due to incompetence and not due to an AI being used to replace their job. With more technologies coming into the picture, the job market gets more competitive than ever, and we owe it to us to keep upgrading ourselves as days go by. But most of the time, at least 85 percent of the people these days (including myself) would rather watch a Netflix movie than go onto the web and watch a useful video about how to develop a Large Language Model or any useful skill that could help them fetch a job or keep one. Our generation has proven to be incredible self-aware, but has also succumbed to more distractions, which stops us from reaching our fullest potential in terms of intelligence, soft skills, and experience.
But how are we going to account for the privacy that we are losing in terms of our data?
Lately, the Ghibli image generation fad has taken the internet by a shock due to its enhanced image generative capabilities. But do people know what is happening in the backend? Not most of them. It is because people these days do not bother reading the terms and conditions or the processes behind some technology before using them. When something is a trend, we just want to jump into it before we know what it is and what it does or rather, how it does what it does. We fail to ask the important questions of what? Why? And how?
Some claim that the Ghibli fad is the biggest scam to steal people’s facial data and use it for unauthorized purposes, causing an issue with privacy and concerns related to it. However, this happens with all social media applications, including Instagram, Facebook or even LinkedIn. Yet, people do not mind posting all about their private lives online. Where is this hypocrisy arising from? While I agree that privacy is important, I strongly believe it is the user’s choice to decide if they want to be on the internet and have their privacy being disturbed or stay off the radar.
So, do I mean to say that I agree with data being stolen and sold?
As a student majoring in AI, at this point in my life, I believe that once we go into the web, whether we like it or not, our data is compromised. That does bother me, but that is why there are classes on cybersecurity and information privacy that we as consumers should have an idea about in case we face backlash due to privacy concerns in our lives at any point in time. For AI specifically, there is a domain of study called AI ethics, which helps us understand our rights as consumers when we use an AI tool. However, since AI is still in its teen years and is not fully developed as a study by its whole, the subject of AI ethics is not as widely known and briefed as Cybersecurity is.
However, the government could do better by monitoring companies, their regulatory policies, and their effect on users’ data. The American government has a whole department known as the Cybersecurity and Infrastructure Security Agency, which oversees and regulates policies related to AI-based software systems.
So for now, the ball is in our court. We can either know our rights and decide to use AI for purposes that do not raise issues over our privacy or not bother using them.
Why do I strongly believe that AI can never replace actual people?
I say this with utmost confidence because something that humans have that machines do not is actual intelligence, our response to stimuli and a physically functioning body which does not require electricity to run. We also have emotional Intelligence and soft skills that an AI agent can never acquire, even after years of training. Furthermore, this training that an AI model goes through is also because of human beings. So, we owe it to ourselves as a race to up-skill, upgrade and evolve ourselves rather than seeing mere machines and tools as competitions. There is a chance given to us to use these machines as tools to develop ourselves; we can either use them or see them as an abomination.
The choice is still ours and always have been.
