Gettin' AI Reet in T'Schools: Na's t'Time

As universities race to make most o' ChatGPT an' them LLMs, they can't let responsible an' ethical AI slip through t'cracks

Image: Midjourney

So, I couldn’t resist posting this—although I suspect it’ll be impenetrable to most readers (sorry!)

I was playing around with ChatGPT’s mastery of dialect (for serious reasons) and asked it to translate a recent post into a speech given by someone with a broad South Yorkshire dialect … in a local pub.

The turn of phrases in the resulting speech (below) are quite delicious. Admittedly I’m a Lancashire lad, but having spent plenty of time living and working in South Yorkshire I’m impressed.

There are a few turns of phrase that I’d change, but not many.

Beyond the rather endearing use of language, it’s intriguing how ChatGPT (using GPT4) is adept at navigating and working with the vernacular of a specific region. I wonder how, in the long run, this will impact effective communication as well as the preservation of local dialects—and their associated cultures.

Of course, I suspect that, for most people, the bigger question is whether ChatGPT can translate from the vernacular to something that is understandable to more than a handful of communities …

The speech below is based on the the recent post “Universities need to be investing in responsible AI now more than ever.”

Gettin' AI Reet in T'Schools: Na's t'Time

As universities race to make most o' ChatGPT an' them LLMs, they can't let responsible an' ethical AI slip through t'cracks

Reet, it's bin summat amazin' to see 'ow fast this ChatGPT, an' all them Large Language Models (LLMs) an' tools, 'ave gone an' changed universities.

Not so long back, we wor all chattin' abaht 'ow to stop copyin' wi' ChatGPT in class. But na, it's all abaht usin' AI advancements in academia to keep up wi' t'competition.

So, this 'ere's turnin' aht to be one almighty whiplash of a change in them schools—an' it's for good reason too. These new AI thingamajigs can change near abaht everythin' we do as academics.

Na folks are lookin' at LLMs as tools for speedin' up discoveries, gettin' them creative juices flowin', sparkin' teamwork, puttin' research groups together, improvin' grant writin', makin' admin work tons faster, an' switchin' up 'ow we learn. An' that's just t'beginnin' of it all, mind ya.

This fancyin' up of AI in schools is partly driven bah fear 'at them early-adopter places'll 'ave a big 'ead start.

The point is, if this 'ere technology can well an' truly make grant proposals better or speed up 'ow new stuff turns into 'igh impact papers or even lead to right persuasive fundraisin', well, them universities 'at ain't usin' it yet will be left in t'dust.

An' wi' things movin' so fast, bein' even a month late to t'party could really mess up fundin', reputation, impact, an' rankin' in t'long run.

For instance, them grants 'at get approved'll most likely be ones done wi' LLMs.

When used reet, LLMs can make sure proposals fit t'bill for fundin', all while bein' clearer an' more convincin'—an' savin' loads o' time between cookin' up an idea an' gettin' it down on paper.

But there's more to LLMs than just that. Good LLMs can 'elp researchers sift through clever ideas an' focus on what's more likely to click wi' funders an' them lot 'at reviews 'em. An' in-'ouse LLMs can sorta capture what a place is all abaht, an' put it all together in a way 'at nobbut few folks or systems can do nowadays.

So, usin' a well-made LLM-based system to bring together teams wi' all kinds o' skills an' spark collaboration in research an' fundraisin' is summat we'll likely see more of.

Imagine if everyone in a school 'ad access to an LLM trained on a massive collection o' documents, notes, publications, articles, CVs, an' all that. Pair that wi' a way o' talkin' to it, an' that LLM could change 'ow we work together, find connections, an' come up wi' new ideas—an' it'd knock t'socks off any system we 'ave na.

That kind o' tech would give any place usin' it a 'uge edge o'er t'others.

Stuff like this is just too temptin' an' important for universities to not be jumpin' on it fast. But as they do, we can't forget to think abaht 'ow to use it right an' be responsible wi' it.

It's not 'bout dodgin' t'problems on purpose, it's abaht not takin' 'em serious enough while tryin' to be t'first in line for t'AI revolution.

Outside o' schools, them folks makin' LLMs, chatbots, an' all them AI things, are dealin' wi' a right tricky landscape o' risks, benefits, an' rules surroundin' their creations. An' these worries are just as important for universities as they are for private an' non-profit groups.

It's them concerns abaht what we don't know 'bout 'ow AI will change t'future 'at led to that letter callin' for a pause on developin' systems more powerful than GPT4, an' 'at's makin' a lot o' places scramble to balance stayin' ahead o' t'AI game while bein' socially responsible.

These concerns aren't daft. They're not based on daft stories or fear mongerin', like some folks reckon. Nah, they show 'ow fast this game-changin' technology is comin' along, quicker than we can wrap our 'eads around it an' work aht 'ow it might affect us all.

Them worries are also helped along bah 'ow eager folks are to get on board wi' these new gadgets, often without thinkin' 'bout what it all means or just brushin' it off, assumin' everythin'll turn aht rosy 'cause it usually does (though we should remember 'oo's writin' t'history of tech successes...).

This, though, is 'ow these newfangled technologies can sneak up on us all an' lead to problems we never saw comin'—not through some daft film-like doomsday, but through folks racin' to be t'first to use 'em an' assumin' someone else'll sort aht any problems.

In all this, them schools 'ave got a big part to play. They not only 'ave t'skills to create top-notch AI systems, but they can also push others to get on board wi' new ideas.

This is an important job for universities, an' it shouldn't be ignored. But wi' great power comes great responsibility, like they say. An' that means askin' serious questions abaht t'impact o' AI on society, what it means to use it ethically, an' 'ow it might affect folks outside o' our own little worlds.

So, as schools are racin' to use AI, they should also be leadin' t'way when it comes to talkin' 'bout responsible AI. An' this shouldn't be separate from what they're doin' on t'inside, but mixed right in wi' it.

There's an old sayin' 'at schools are good at tellin' others what to do, but not so good at takin' their own advice—'specially when it comes to doin' right by society.

We can't afford that wi' AI. We never could, but t'stakes are sky-'igh na. That means, as schools rush to use AI in research, admin, learnin', makin' a difference, an' all sorts, they need to be investin' in makin' sure AI's used right an' is fair to everyone. An' they need to work together wi' other folks facin' t'same challenges.

Not doin' that just makes 'em one more part o' t'big wave o' AI change 'at thinks bein' responsible is important, but leaves it to someone else to figure aht.