Colombia's passive examination system

@wwaycorrigan

[For an audio/vlog version of this story, click here.]

A few years ago, a Colombian ex — of sorts — asked me for some urgent help. She had an English exam to take for her university. If she didn't pass it, she wouldn't be able to graduate.

She had one slight problem, though. She had very little English. Well, she practically didn't speak the language at all.

'The exam must be passed; by somebody, it doesn't matter who.'

Not-very-testing times

Cue a couple of weeks of intense English classes dictated by me to get her up to scratch. That, of course, would have been the more righteous approach. What really happened? I took the exam for her. She still, rather pathetically, had a hold over me at the time. And she knew it. At least she did watch me answer the questions; I'll give her that much.

The way that text-only, multiple-choice exam was set up — fully online, unrecorded, and with a very loose student verification mechanism — it seemed as if the university didn't really care if it was taken by the student who was meant to take it.

In other words, it effectively encouraged cheating, particularly in a country where the overall level of English, as well as the desire to speak it, is low.

Around that time, I did write a piece, Colombian universities' English at all costs, questioning the requirement for a certain level of English in courses where knowledge of the language was far from essential.

'It's not that the student must pass the test. It's a case of — using, pertinently, English's passive voice — the test must be passed . . . by somebody, it doesn't matter who.'

But eight years or so have passed since that old flame asked for my assistance; plenty of time for the universities and language institutes to re-evaluate their procedures and amend them where necessary. Except, from my latest experience, they haven't.

Sound system

For one particular institute, in any case, the acceptance of cheating — or at least playing the system — is even more blatant than what I'd witnessed before.

Here's what happened: An 18-year-old who I've never met — he was given my contact details by a production assistant on a shoot in which I recently participated — asked me to take his test for him.

This time, however, it wasn't a text exam. It was an oral one. Yes, an interview via a Zoom call. Thing is, it can be done without using the camera. The examiner just wants to hear the student's voice.

Thus, with my, um, Mid-Atlantic English accent — or is it still a west of Ireland one? — I assumed the identity of this Bogotá native teenager.

Now, if the one conducting the interview — Laura, a Colombian, who was visible to me throughout the call — genuinely thought that I was an 18-year-old studying English as a foreign language, either she's not the sharpest examiner around or my level of English is not what I thought it was. Or maybe I'm just a good actor, although I made no effort to come across as a rookie learner of the language, as the guy for whom I took the test clearly is.

Perhaps, seeing as how it must have been obvious to Laura that it wasn't an 18-year-old Colombian answering her questions, the oral exam will be declared invalid.

However, from what I gather, it's an accepted practice in certain quarters.

In essence, it's not that the student must pass the test. It's a case of — using, pertinently, English's passive voice — 'The test must be passed — by somebody, it doesn't matter who.'

Deep fake

Yes, I could have declined to take it for this guy whom I've never met. For one, the payment I received for doing it won't tempt me to dine in an exclusive Bogotá restaurant. It's barely enough for a roasted whole chicken in my working-class barrio.

Yet, curiosity got the better of my mild moral objections to this rather innocent fraud. If I hadn't done it, somebody else would have, no doubt.

What's more, if the guy was more tech-savvy, he probably could have set up ChatGPT or the like to do it for him. I'm quite behind the curve when it comes to the use of Artificial Intelligence, but from what I've heard and read, it would be up to such a task. In fact, I imagine an AI version of the student would have been a far more convincing fake than me.

Thus, with AI's overhauling of innumerable areas of human competence, Colombia's passive examination system is perhaps a better reflection of the times than what's considered best practice in other places.
__________________________________________________________
Listen to The Corrigan Cast podcast here.

Facebook: Wrong Way Corrigan — The Blog & IQuiz "The Bogotá Pub Quiz".

More artificial intelligence, less human intelligence

@wwaycorrigan

[For an audio/vlog version of this story, click here.]

While not officially a mechanic, my father is fairly well-versed in the workings of the combustible engine, its accompanying components and other such machines.

'You can leave your brain at reception. It's not needed here.'

In the rural Ireland of his youth, the sight of a car was a memorable event. Thus, being more technical than academic — the latter wasn't really open to him in any case — he was drawn to discover how these mesmerising motors operated. From what I've witnessed, many of his peers from similar backgrounds did likewise.

In this earlier phase in the rise of the machines, it seems the desire to understand the transformative technology was widespread.

That the rate of change was slower and physically more in one's face, so to put it, this inclination to educate oneself about the new developments made much sense. Even Luddites had a decent grasp of what they were opposing.

AI kills the radio and video stars

Today's more rapid change, occurring in what is a largely intangible ethersphere, even for netizens, has left many of us in the dark about the workings of the digital devices driving our daily interactions.

When problems arise, as they invariably do, one is unlikely to have a DIY solution, outside of a reboot or the more drastic factory reset. And, as most will know only too well, many digital difficulties, online obstacles, require a fix far more convoluted than an unassisted system restart.

In my November 2022 piece, The digital dystopia, I wrote about some of these tech troubleshooting travesties.

'Large language models aren't quite at the stage where they can churn out Churchillian-style rhetoric but, by all accounts, they can come up with bread-and-butter prose satisfactorily.'

Now, with the inexorable advance of artificial intelligence (AI), not only are many of us ignorant about the nuts and bolts that put it together, but its raison d'être is to do the majority of thinking for us.

Rather than go through the hassle of doing strenuous study, racking your brain, AI will rake the internet for you and come up with a satisfying solution almost instantly.

So on the artificial face of it, life will be less arduous.

In my line of work — if I can be said to have one — the need for human editors and proofreaders will greatly diminish, nay already is diminishing.

Grammarly and the like have been aiding writers for the last number of years. Such grammar software is far from perfect but it's improving rapidly. (As Grammarly currently is, I wager that it's better than the average native English speaker at spotting grammatical errors.)

If Grammarly is taking the place of the real-life editor or, worse still, the need for one to grasp basic grammar, ChatGPT and its competitors are doing away with the need to be able to write in the first instance. Again, these large language models (LLMs) aren't quite at the stage where they can churn out Churchillian-style rhetoric — who amongst us can? — but, by all accounts, they can come up with bread-and-butter prose satisfactorily.

I imagine that LLMs will do — perhaps already are doing — a lot of the heavy lifting in compiling TV/radio news bulletins and suchlike. Heck, this might even tempt me back to the newsroom! (Refer to my piece Making the bell toll for us while we still can for more on my previous newsroom nightmares. The DJ-3000 from The Simpsons episode Bart Gets an Elephant, first aired on 31 March 1994, also comes to mind here!)

As for learning a foreign language, this could become obsolete save for those who have a special interest in it. Surely, in the very near future, Google Translate will be viewed as a caveman-like tool for communicating in a non-native tongue. 'What, back in the early 2020s you still largely tried to speak a foreign language?'

Active intelligence

With all of these AI tools, one can try to be an active, inquisitive user. I mean active here in the sense of not merely accepting uncritically the AI answers. For example, coming back to grammar, if it suggests a correction, one should at least know why it's doing so.

In this way, one can also learn. The rise of AI doesn't have to signal Armageddon for humans — or at least Armageddon for the use of one's critical faculties.

While engines and motors rendered certain lines of employment redundant, it can be argued that they paved the way for today's health and fitness industry. Work for many of us may be more sedentary now compared to previous generations, yet we still seek physical stimulation — be that simulation or real. And machines have certainly helped bring about mass tourism, problematic as that has become all the same.

Similarly, AI comes with opportunities and challenges. Some academics and experts fear it has far more of the latter. For sure, it is set to be more intimately disruptive in the lives of every one of us compared to previous technological advancements.

With that in mind, should refusing to engage with AI become next to impossible, trying to ensure that we remain in control of it rather than it controlling us, is key.

Keeping well informed and as much in the know as possible is a start. The overall direction of travel, though, suggests we're entering an AI Bright Age, with a new Dark Age for human intelligence.
__________________________________________________________
Listen to The Corrigan Cast podcast here.

Facebook: Wrong Way Corrigan — The Blog & IQuiz "The Bogotá Pub Quiz".