It seems that I have been experiencing a continued decline in the accuracy and reliability of the AI tools I use. I generally like Perplexity and have used it for tasks like summarizing documents, reviewing interview transcripts, editing, etc. I always double-check results for accuracy—and that’s a good thing because I’m finding these tools to be increasingly unreliable.
One of the big downfalls of GenAI is that it currently only draws from content it is trained on that already exists. So, the information it generates is “known.” GenAI can’t (yet) create new material; it can only consolidate and spew back what’s already been created by others. In other words, “it” cannot be innovative.