Mini-Vent: The New Trend in Hiring - AI Skill-Assessment Testing.

 

So I'm applying for a second job, right?

I love my current job, but I'm a contractor and the work has been slow, this year, which I'm assuming is due to the Inflation Reduction Act changing the game for pharmaceutical prices.

Now, I realize that potential employers could read this, and yet I'm fine with that, hoping that they'll take this as feedback for those trying out this new trend.

So taking tests is almost always a prerequisite in medical transcription, but the typical MO is the employer creates, administers, and scores their own testing, and I usually ace them.

HOWEVER, there's this new trend where many companies are using either Indeed's skill-assessment testing or a third-party test assessment, but regardless, both are usually created and scored by AI.

(Well, there was one that I took that may have been created by a human, because it was thorough and accurate, but still scored by AI.)

Now, if you take them through Indeed, you cannot take them again for another six months, and there's no warm-up.  

Thank God I scored proficient on all.

However, this morning, the medical-transcription knowledge test that I took was administered through a third-party service for a company and it was very thorough - and I failed it.  

 (Mind you, I have never failed one of these tests in my life in 26 years and I actually thought I scored very well on it.) 

Thus, Indeed automatically marked that job in my queue as the employer having declined me automatically.

Now, I knew I was a bit rusty with true, by-the-book AAMT BOS  medical transcription guidelines, as it had been five years since I'd done hospital transcription, but WTH - am I getting early dementia or what?  😂

I became so discouraged that not only did I feel like giving up on finding a second job, but I started wondering if I even deserved the med-pharm job I have now that I've been doing for five years! 😂

HOWEVER - imagine my surprise when the recruiter for the company wrote me that I could proceed with the audio portion!

She said the answers that AI marked wrong were actually correct!

Oh, thank GOD, I'm not losing it after all!

They apparently were testing out this new testing and wanted to see how it went, we were guinea pigs 😂. 

So it asked me things like to correct the spelling for "drug-eluding stent."

Thus, of course, I wrote "drug-ELUTING stent" in the blank, which is correct.  

However, AI marked it wrong, and the problem was, the test wanted me to correct the spelling of just "eluding."  

However, I had written "drug-eluting" because it's a hyphenated modifier and therefore is considered one word!

(Mind you, the instructions said just "correct the spelling.")

Other stuff like that, which apparently still confounds AI, it still thinks very binary, when you just can't in some fields 😂

Thus, the recruiter found that I had actually gotten all five answers correct that AI had marked wrong, so I proceeded with the audio-transcription test, which they create and human-score themselves, which I completed and am waiting to hear back.

Ya know, considering transcription is rapidly being replaced by AI and we all live in fear of it - prematurely, I might add, over things like this - you'd think they, of all people, wouldn't have even considered AI testing and scoring, right? 

I guess the lesson we all learned here from them trying these services out is that all our jobs are still safe, for now  😂

But the one I just took tonight takes the cake!

So it was another test from another third-party testing service for a medical transcription position.

Did it have anything to do with medical terminology, anatomy and physiology, or even typing, spelling, grammar, Microsoft Office familiarity, or clerical stuff? 

No, it did not. 

It was essentially an IQ test, or more accurately, a college entrance exam like the ACT test, except instead of being given appropriate time, you're answering as many of 40 questions as you can in 12 minutes!  (I completed 24, hopefully correctly, although the last one is iffy.)

Things like ..."Unscramble the following letters to create a word that means "confident" - DRESASU.  What is the last letter of the word" - and then you chose multiple choice for the letter "D" being the last letter of the word "ASSURED."  

(That should actually be "self-assured" to actually mean "confident,"  if we're splitting hairs, not just "assured," but such is AI testing and it was the only word you could create out of the scramble that made sense to mean confident.)

Or like, "Planning is to goal as injury is to __"  - A) pain, B) winning,  or C) poor training - the answer of course being C, poor training, although this isn't always the case in the real world, but out of those choices, that's the ticket.

High-school math, and stuff like "Finish the sequence:  24, ___, 29, 32, 36 (answer 26),

Then some logic questions, like "Some people aspire to be promoted.  Most people believe they should be promoted.  Therefore, most people that believe they should be promoted, are promoted" and judging whether that is a valid argument or not.

(This is NOT a valid argument, of course.  There is not enough information in the logical argument presented to ascertain whether people are actually promoted.)

Keep in mind, you're being timed - you have 12 minutes to answer all 40 questions.  

I answered 24, hopefully correctly, but I don't know because this one didn't tell you your results!

WTF, am I trying to get into Harvard or just trying to get a clerical job that pays under 50K a year? 😂

I mean, FFS, it's a clerical job, just transcribing a doctor's basic dermatology office visits!   😂

So this particular test didn't give me the results (it just said "Thank you for completing the assessment," so I have no idea what I scored on it. 

Now, my IQ has been tested 5 times, and it always falls somewhere between 120 and 129, somewhere between above average and superior, 3 times at 125, so I go with that, if asked.

 (I scored the highest on the MENSA test at 129, but you have to be 130 and above to be in MENSA, so it was a no-go).

But frankly, even if I scored genius level on the thing, why would my employer need to know my actual IQ number?!?

That's pigeonholing, and not to mention, kinda creepy, right?

Because let's say you score genius level or aptitude - are they going to expect you to be able to solve all of the practice's problems even better than the doctor?

Are they going to pay you $100,000 a year, just to type office notes for a dermatologist, just because you scored genus level on this test?

I don't think so.

Or let's say you're the sort of person who isn't a good timed-test-taker because you're dyslexic or you have ADHD or have severe test anxiety or something, or that you were coming down with the flu when you took it, so you scored lower -  are they going to talk to you like you're mentally disabled and say you're only qualified to like, clean the bathrooms or-?  😂

Also, I'm not sure a 40-question test in 12 minutes could even determine that accurately, nor do I think that we should be automatically equating speed with intelligence anyway.

So out of curiosity, I Googled this new hiring trend and found several Reddit threads where people were really upset/frustrated by it.

Though there ARE "highly proficient" and "expert" ranges - no one in any of the Reddit threads tht I read had ever scored either of those, just "proficient," including myself  - though some didn't pass at all in their fields, despite degrees in them.

In fact, I saw two comments on Reddit, an electrical engineer with a master's degree and an aeronautical engineer with a PhD respectively, both scored "proficient" rather than "highly proficient' or "expert" on the Indeed-administered 20-question engineering tests! 😂

The PhD in aeronautical engineering actually withdrew his application later because he felt it was insulting and demeaning, to first of all be tested at all when he has a PhD, has published white papers, journal articles, and books, but even worse, some AI 20-question program that has no real grasp on the subject matter? 

So in the Reddit thread that I read, it seems most employers don't take these overall seriously, it's more of a pass/fail than anything.

Regardless, let us hope this trend of quick-shot testing is found to not assess pretty much anything accurately and is NOT a foolproof way of eliminating or promoting candidates, and thus dies quickly, yes?

I feel sorry for skilled employees that were actually ruled out because of this flawed process, right?

Regardless, scramble this - FUAI.  I hate U already, for a multitude of reasons!  😂 😂 😂