Trego's Mountain Ear

So I see a facebook posting that lists Murphy’s Law as ‘the more you fear something, the more likely it is to occur.’  Would have been a surprise to Murphy, who said “If anything can go wrong, it will.”  Sam sent me another AI generated theme that claimed “No person, no problem” came from the film The Lion King. 

Pretty sure that phrase came from the great Soviet philosopher Joe Stalin.  Heck, Uncle Joe left us photos to prove it:

The problem I’m seeing with artificial intelligence is that, while AI can quickly review the literature, it lacks the ability to separate bad data from good.  I’m not seeing announcements on AI making discoveries that alter our understanding of physics.  I am seeing where high school students and undergraduates are using AI as a way to plagiarize papers.

It’s a fast way to do a literature review – and that’s essentially what a high school term paper is.  The problem is discernment.  The cop instructor at TSJC and I were about the same age.  He had completed a cohort based master’s in public administration – I had a research bachelor’s in Sociology and had then gone back in Ag Engineering.  He started a doctorate (pretty much correspondence) from Nova – and was shocked to learn that including the Encyclopedia Britannica in your sources wasn’t acceptable in a Ph.D. program.  His master’s degree was one of those termed a “terminal degree” – meaning that it wasn’t meant to prepare him for a doctoral program.

And that seems to be the weakness of AI programs – they can definitely read faster than I can, and condense the information – but AI (so far) doesn’t have the ability to sort out the bull.  The spot where AI scares me is that students will use it, rely on it, and not have the ability to discern bad research, bad assumptions, from good.  I can see where AI has great potential for speeding research – but I remember my own dissertation, where a single conversation with one old Hutterite minister disproved my hypothesis.  When I knew what I was looking for, I could find the data – but without that conversation I would have published bad research. 

My belief is that AI can develop some very sloppy research habits in students – and it will probably be more damaging in my own area of social sciences than it will in quantum physics.