The End of the World?

By now you must have heard of AI, or Artificial Intelligence. But do you fully understand AI? Hopefully I can help you grasp the pros and unfortunate cons with Artificial Intelligence.

What is Artificial Intelligence?

Let’s start with defining AI. Artificial Intelligence is a computer simulating human intelligence in a more efficient manner. It learns human intelligence by consuming large amounts of data put into their training modules to create correlations and patterns. This can in turn make them predict future conversations.

Think of AI as a human baby. When a baby is born it knows very basic functions to survive. How to breath, blink, rest, swallow, and of course cry. You as a parent must teach the baby how to grow further. For example, you put the baby in front of the father and get them to say “Daddy”. You say it enough and the baby will corellate that this man in front of them is “Daddy”. A baby will also watch mouth movements and the tone of your voice to determine sound, or patterns from repeation. Once a baby learns how to speak from listening to you parents yap all the time, they can correlate the sound and name “Daddy”, to the man in front of them. AI acts very similar to this although with the benefit of being infinitely faster with analyzing data input.

AI programming focuses on cognitive skills similar to the way us human do like:

Learning – This part of AI programming focuses on gathering accurate information and then how to turn it into actionable information, based on the request. Algorithms provide a step-by-step instruction on how to complete tasks.

Reasoning – The AI learns which algorithm to choose from to reach a desired outcome based on the specific request.

Self-correction – This part is very important to an AI algorithm. This aspect of AI programming focuses on tuning the algorithm each time a desired outcome is not favored by the individual making the request. If I ask Chat-GPT to make me a healthy low-calorie recipe and it provides me a menu to McDonalds, I will express to Chat-GPT that McDonalds is not healthy. It may or may not come back with something more accurate. We will get to this later.

Creativity – With all of these focuses brought together, AI algorithms can then base all of this correct information of pattern, reasonings to a desired outcome, and self-correct to fine-tune the mistakes to create its own images, text, music, and overall new ideas.

To summarize, AI learns from large amounts of data input. This may be from ingesting the entirety of the information of the internet, or new input from users like us. Then when we ask AI to do something, it will take the data it has gathered to find patterns to create a desired outcome. If the desired outcome from the AI model does not suit the user, it can self-correct itself to find a different outcome. The more data input that the AI model receives, the better it can be at creating its own ideas. Now that may sound scary but let’s delve into that sentence.

The Limits of AI

Artificial Intelligence has become a part of many people’s everyday lives. I catch myself asking Chat-GPT more questions than Google. Most individuals use AI in the workplace in jobs such as marketing, graphic design, and Information Technology. I get it, why would I create a super long html code to create a website when Chat-GPT can create it a million times faster. Although, at some point we have to wonder what the limitations of AI are. AI is still a computer after all, and computers are dumb.

These limitations below outline the drawbacks to fully relying on AI in our everyday lives:

• AI does not understand the nuances of the human language. AI models still struggle to identify when someone is using sarcasm, irony, or figurative language. This is different from saying something sarcastic to saying a phrase that is commonly used as sarcastic as “It’s okay if you don’t like me. Not everyone has good taste.” This phrase is used only in a sarcastic tone and AI would know that. If you reply “Thanks, you are so very smart” to a clearly wrong answer, the AI will acknowledge the thanks and move on.

• AI is terrible with common sense. Which does make it harder to have a genuine conversation with it. If you train an AI model to make bad decisions like underage drinking which is just dumb, a teen might use the AI model to ask, “Should I go to my friends party, there will be alcohol and I’m scared to get in trouble”. The AI model will pull up some statistic on how often underage drinking occurs and nothing has happened to most people so they should just go. AI bases it’s output on the training data and human input only. It cannot apply the knowledge it may gain from others saying underage drinking is a problem and not good advice.

• AI has no way of expressing emotion. This is not to be confused with expressing the outcome as emotional. By knowing the definition of a jealousy, anger, depression, anxiety, and happiness, does not mean that AI can act those ways. AI only processes data as logical and in a structured way. AI will always speak to you as they are fixing a problem, and not as someone who just wants to be listened to.

• AI can be biased on the desired outcomes based on the data they were trained on. For example, if I create an AI model to help me determine who should be on my fantasy football team but only include players in the AFC East in the training module to analyze stats. I have a feeling that all other players in the rest of the league will not be included.

The short answer is no. Although you should not completely trust everything. Malicious people around the world have used AI to create stronger hacking campaigns using voice and video impersonation, phishing email creation, and new algorithms for password cracking. AI is far off from becoming sentient and there are many other factors that would make it impossible as these advancements haven’t been invented yet. You can and should use AI as a tool and resource. The more you play with it, the better you can understand it. Once you interact with AI, I promise you will understand that it is not smart enough to take us over…or is that it’s plan?