illustrations by StorysetEarlier today, my wife asked me to review a letter she wrote. After going through the post, I prompted ChatGPT to review the letter, filtering grammatical errors and removing redundancy. However, my wife had an issue with this process. When asked why, she couldn't give me a precise reason, and I could not persuade her that the text was hers. This story highlights the relationship or perception most people have about Artificial Intelligence.
Most people view Artificial Intelligence as a direct competitor. This can be attributed to the negative press and fearmongering around Large Language Models (LLMs) like ChatGPT, which peaked during the AI hype last year. People were sold the narrative that AI will take their jobs and eventually replace them in every sphere. So, people are justified in their fear because they do not know any better. Fortunately, my experience with LLMs is a lot more positive, and it is something I use daily.
At first, I was indifferent about the AI hype. To me, it was just another marketing gimmick to get people to part ways with their money. After some time, I got accustomed to LLMs like ChatGPT, but they were also demonized after some time mainly because people were using them in wrong ways. Fortunately, that is not an issue with the tool but with those who use it. Just like crypto, you cannot disregard the benefits of a thing solely because of bad actors in a sphere.
As a writer, LLMs have helped me a great deal. Previously, I had the onerous task of writing and editing my work. I sucked at the latter. I always needed an extra pair of eyes to review my pieces for grammatical errors. As I could not afford an editor, I relied on friends and family, and this made my creative process a lot slower. Now, thanks to LLMs like ChatGPT, I have a personal editor that helps me make my posts a lot better. This has improved my efficiency and output over time. A post that would ordinarily have taken 3 hours to write and edit is cut down by half, and it does so without taking the soul out of my piece.
WHAT THE CRITICS SAY
One of the criticisms many people (like Paul Graham) have about LLMs like ChatGPT is that it takes out the soul of a write-up. People have their quirks, and those elements of their being should reflect not only in speech but also in writing, and I agree with this idea. However, I do believe in effective communication, and this can be bolstered by clear and correct messages. This is a role LLMs can fill at little to no cost.
Now, the problem is most people are not good at creating prompts or would rely on LLMs to generate content for them. This is lazy and destroys the creative process, and I totally frown upon this use of AI. Humans need to work with AI to improve communication through concise, correct, and complete messages.
WE ARE NOT THERE YET
illustrations by StorysetIt is worth stating that AI isn't at the level most people fear. For instance, one criticism AI gets is that it creates dependency, thereby turning people into human zombies. My rebuttal to this criticism is that humanity stands to gain more from working with AI. As long as this technology is accessible to everyone, it stands a better chance to improve lives and create a level playing field for those who do not have the means to compete fairly in the world today. This cuts across industries like education, agriculture, medicine, etc., but we are not there yet.
For instance, Marcus BrownLee is in a saga over his review of the Human AI device. This is one of those cases of a founder trying to take advantage of the AI trend to create a half-decent device to sell to the public. In BrownLee's words, "the tech is way ahead of its time." Many people do not understand the limitations of LLMs, which grossly skews their perception of Artificial Intelligence.
There is an interesting podcast on Spotify by Lex Fridman featuring Yann LeCun that extensively exposes the limitations of LLMs. It clears some of the misconceptions most people have about Artificial Intelligence. I believe conversations such as this are more wholesome, providing people with a clear picture of what AI is and could be in the nearest future, and the part humans play in this technological evolution.
CONCLUSION
Artificial Intelligence is here to stay. It will only get better over time. Fearmongering would only stop people from using this technology, which puts them at a disadvantage. We must encourage people to use these tools to improve their work and communication channels.