In the last few weeks, there’s been a vibe shift. And, no, I’m not talking about Kamala Harris.
GenAI, or any AI-created output, is suddenly being described as meh, if not downright off-putting.
Customer service departments who’d integrated it into their operations are now getting pushback from their clients, who simply want to interact with an actual human.
On LinkedIn many are growing wary of reading AI-generated, or at least AI-assisted posts, and what are, to some extent, AI-generated comments. Perhaps more often than LinkedIn users would like to admit, AI-generated posts are being responded to with AI-generated comments. And the button inviting you at the bottom of each post to — “Rewrite with AI” — only makes AI-generated prose more common.
In some instances the stakes are even higher. According to a recent Gartner study, healthcare and financial service industries are getting significant pushback from their clients who, as soon as they think they are engaging with an AI, feel put off. Indeed, a majority of these customers were found to have major concerns about AI in customer service.
In short, people have become fed-up with AI, especially the bland and generic kind of writing that has become associated with AI.
To many, these recent developments might sound like we are well over our AI crush. The reality, however, is that AI is an ever-evolving technology. Today’s generic reams of prose will become ever more refined, and probably in a very short time. To illustrate, in the last few months, many have begun favoring the use of Claude over ChatGPT, citing its ability to write better–more naturally, more eloquently–than ChatGPT.
While individual users can quickly jump ship for the latest, shiniest models, enterprises are often locked into whatever proprietary model their developers built (often to protect an enterprise’s data, as well as train it on an internal corpus.) Much of the bad writing we are seeing is from ChatGPT 3.5, which, while relatively advanced when it stormed to 100 million users in 30 days, is still painfully AI writing, nothing like the more eloquent writing of Claude’s latest model–Opus 3.5.
Once an even more refined model comes out, and enterprises not looking to alienate customers (but still looking to reap the operational savings that potentially come with GenAI) start using the latest models, our kvetching over being inundated with AI speak will likely dramatically decrease.
And in all likelihood, most of us will not think twice about using these latest models for the majority of–if not all of–our writing.
So AI is not going away, and it’s only going to get better. We are simply getting over our first infatuated throes, becoming a little more critical of the substandard writing that has proliferated around us with the inception of 3.5, and are demanding for something more human-sounding.
However, whether that actually comes from a human or an AI, might not be that important.
Disclaimer: every single letter of this piece was written by me, with absolutely no interaction with an AI.
Comments