Experimenting with AI Text Generation

With AI technologies at the forefront of our modern society, the people have an array of questions. As I'm writing this blog post, I have the option to utilize AI to assist in generating a full length blog post, a compelling title, a structured outline, or even meta tags that can significantly enhance search engine optimization. Some find that using AI while writing can help streamline the writing process but also spark creativity.

The discussion this week is centered around a critical topic: whether or not we can place our trust in AI when using it in regards to factual information. This is an increasingly relevant concern, especially as AI systems become more integrated into our daily lives and decision-making processes. To explore this issue further, I decided to conduct a small experiment using ChatGPT to complete a prompt for me about Vale Tudo, an old school form of combat sport. I specifically asked ChatGPT to, "in 300 words, explain what Vale Tudo is." The results left me genuinely impressed, showcasing the potential of AI in generating coherent and informative content.

The resulting blurb that was crafted for me was not only entirely accurate, but it also managed to encapsulate the essence of Vale Tudo in a succinct manner. However, I noticed that it left out some crucial details that are essential for understanding just how captivating and different these events were. One significant aspect of Vale Tudo that made the bouts particularly shocking and enticing was the complete absence of weight classes. This meant that a fighter weighing 200 pounds could face off against an opponent weighing just 125 pounds, creating an unpredictable and thrilling dynamic. The outcome of these matches hinged not merely on size or strength but rather on fighting style, technical skill, and the individual fighter's ability to adapt to their opponent's tactics. Such a setup not only tested the limits of physical prowess but also challenged the mental fortitude of the competitors. Feel free to take a look at my Google Doc here, where I edited AI's result.

From this experiment I realized that you can't fully trust AI to write comprehensively or accurately. AI reminds me of Wikipedia in the sense that it can be used as a good baseline to spark your own writing, but it shouldn't be used as a source itself. I would say I trust AI about 50% because I know it is sourcing knowledge from other sources, but those sources aren't always accurate and reputable, which is where the inaccuracies come in.