Today, I found out that working with generative text AI is extremely difficult. I hated every last moment of it. I honestly don’t understand how some people actually use this as a tool to help them with assignments.
First of all, the blog post it wrote for me from my prompt was generic and bland. It didn’t seem robotic, but it did come across as if it was written by the most boring person who barely cared about the topic. Second, it confused quite a few things as actual facts because there were similar relevant scenarios. For example, it confused years when my subject left and rejoined a group. Finally, in order to not outright copy a bot, one has to basically rewrite anything it spits out at you before you use it. This was so time consuming! I’d much rather just write the whole thing from scratch by myself without its help.
As far as how much I trust this software (I used ChatGPT), the short answer is I don’t. Despite its common descriptor, there is nothing remotely intelligent about generative AI. It’s basically a pattern completing software with a huge database to pull from. I’ve worked as quality control evaluating AI output for major companies. They can try to get it to make better patterns, but it will never understand the input or the output well enough to make it worth the cost in terms of the environmental impact.