Recently I experienced a foolish moment in my life. I suspected someone using ChatGPT for their work, in this context weren’t suppose to use it. Initially I though the numbers were off based on my own calulation and the write up was done immaculately. I suspected they use chatGPT because it was done so well, it even had equations that chatGPT spits out. I never see such clean equations and the writing and sequence reminded me of chatGPT outputs of equations.
I meet with the person 1 on 1 and asked could you walk me through your calulcations and how did you do the equations? They did it all perfectly in front of me and foolishly I was the one to suspect them, I "Jump the gun" so to speak. I admitted my assumptions and apologized at the end of our meeting, I was clown to suspect others and wrongly suspected my fellow. I became one of those people who I couldn’t tell the difference between humans and chatGPT.
I suspect things of others and became so cynical. I have been told that I was too gulible. So I have tried recently to be more skeptical. The problem with skepticaism is that makes you blind to potential opporuities but it also protects you.
The Death of Truth
However, there is a worry that I have in the back of my mind about ChatGPT and its generative applications is that it can so easily generate data points, and write convicing reports that we are going to be in a ocean of misinformation. Of course we have been experiencing that already but imagine on a scale much worst drug data, and engineering data, could be falisfied.
if you are not fimilar with scientific fruad, well its actively happening and its been caught by scientific slueths such as Elisabeth Bik who catches and identifies this fraud where they are AI-generating images and putting them into scientific papers! You can read more about her work at Science Integrity Disgest.
The Rise of Scientific Reproducability
Now that were witness the rise in fraud within all information and knowledge systems. I suspect we will need a industry to be contracted as 3rd parties to audit and validate those results from scientific studies and engineering projects. Even 3rd party auditors are suspectiable to bribery such as the Volkswagen’s Diesel Emissions fruad Because so much of our products we use today completely failing quality checks and reproducabile results that its cost us millions and even our lives. All iinstitutions today have to be as rigourness, if not more when it comes to reproducability and validation its cost that society that doesn’t want to afford as whole but will only pay later in lives.
We will always be triple checking each others work until the end of time now. I guess that is good thing.