Ok this is fascinating.
Some of you may have heard of the Replication Crisis plaguing academia.
In short, most scientific studies are, well, bunk. As much as it pains us all to admit, that's just the case. Not necessarily wrong, per se, but invalidated by the fact that either the vast majority (less than 25%, sometimes as low as 10%) of research is reproduced, either because there's no incentive to retry the research, or if you do retry it, the results don't match the original.
So yeah, this is a known problem, although it is technically part of the scientific method, a journey of eternal forward motion, improving and zoning in on truth. Some things are necessarily corrected or debunked on the way.
But hold your horses.
Don't you think for a second that things are getting better. We are humans after all! We're not going to stand by and let our brains and culture set things up for success!
Why bother doing that when we can simply imbue more chaos into the ecosystem?
Let's Delve Deeper
I like Google Trends. It's fun. You can see how words change over time either with social media influence or just cultural shifting. You can pinpoint specific events in time simply by observing when a particular word breaks out in popularity. For example, here's Trump:
Here's 2021, 2022, 2023 and 2024:
I wonder what happened in November 2022?
Similarly, here's an interesting trend. See if you can figure out what's going on:
Delve?? Why delve?
Any guesses?
Turns out, this is a result of scientists using CHAT GPT in their papers! ChatGPT apparently uses certain words more often than others. I'd feel sick if I wasn't so jaded and unsurprised.
Is this going to matter? Is the fact that scientists are already highly dependent on AI to do their dirty work going to destroy the very foundations of the scientific method? Are we really that lazy as a species? Or is this going to be no problem whatsoever and it just saves some people much needed time?
Well, the use of AI in writing out research is a whole thing someone else can delve into, but the consequences to the greater society is what gets my mind tingling.
The more one creates works that trigger ChatGPT to use delve, the more the algorithm uses delve. The more people notice it, the more people talk about it, talk with it, which then creates a feedback loop where delve grows and grows.
Eventually, delve is going to, if not already, feed the zeitgeist of common parlance, and it becomes a far, far more common word than it once was.
Our very language changes dramatically.
It's too early to Google Trend this particular instance to see if delve gets used more in every day online life, but eventually we might indeed see this change over the years.
AI, like social media, could become a steroidal boost to the manipulation of our language. I mean, the very nature of our language is an organic, fluid mess, subject to change on a constant basis. But I do wonder if this is going to accelerate change from centuries down to decades.
Will people 100 years from now be totally impossible to understand as English hyperdrives its way through its own evolution?
In Japan, an old man once sued the public broadcaster because they used too many borrowed words from English. The elderly simply couldn't understand what was going on around them anymore, the news was becoming total jargon to them.
Linguistic engineering can go pretty far. In China, they are known to go through great authoritarian means to prevent such influences to their language, despite massacring the language themselves mere decades ago by creating 'simplified' characters.
Other languages simply go extinct. A full HALF of current, living languages may be going extinct within the next century, as we homogenize the world with broad brushes of English, French, Hindi and Mandarin Chinese. There's 26 languages in France alone that are set to get wiped out, in place of, well, French.
That's a really, really rapid change.
Well, I'm not sure what direction this blog was going, but I figured it was pretty interesting.