OpinionPREMIUM

TOBY SHAPSHAK: AI won’t save us from our bad habits

Shortcuts lead to long-term setbacks

AI isn’t making people more efficient; it’s just making us lazier, says the writer. (Dado Ruvic)

AI may change the world, but it will do nothing to fix our own bad behaviour, bad habits and prejudices.

First, you need to remember that AI isn’t actually intelligent. Yet. It performs some tasks very well and is learning to do more useful things for us. It’s very good at summarising vast amounts of information on the internet.

However, if it can’t find an example to back up its précis (the fancy French word we learnt for it in grade 6), it just makes it up. This includes court cases and legal precedents, as numerous embarrassed lawyers have found out to their shame.

Second, it’s a parrot. It repeats what we say or ask it. Trash talk an AI chatbot and it trash talks back. Witness the legitimate furore about Elon Musk’s Grok AI chatbot, which was allowed by its developers to accept a prompt to undress people. AI did the horrible thing, but a human enabled it first. D’oh, as Homer Simpson would say.

Educational setbacks

Third, shortcuts are just that. People who use generative AI to do their work or school assignments are unwittingly fooling themselves. Research has shown that university students who use GenAI to write their essays are facing a three-strike problem.

By not reading through all the material yourself to summarise it, you’re both not learning how to summarise and you’re not retaining any of that information. That is a crucial step in the education process.

Next, those short-cutters aren’t learning the equally essential skill of discerning what is important from what isn’t and how to rank the importance of information.

Then, the final whammy, those students just don’t remember either the subject matter or learn the education process. They used GenAI for so-called cognitive offloading, produced the end product and ended up with nothing. It’s self-destructive in the long term.

‘Empty calories’

GenAI tools are the empty calories of education. We will live to regret unleashing them on our kids, as much as we have smartphones and how they have destroyed a generation’s attention span.

AI isn’t making people more efficient; it’s just making us lazier. The signs are everywhere. You can tell an AI-generated copy at a thousand paces even if you aren’t a professional writer. People are using GenAI services and producing frankly poorer results.

The hype about GenAI’s ability to rewrite copy and summarise does not match the result. It’s not just the blandness but that the writing has no soul.

What is most worrying is that the people who use it are so enamoured with it that they fail to notice how lifeless and soulless it is. It’s like those posh homes you see on TV that are decorated by an interior designer and never look like anyone lives there. It has no soul.

I am an expert in identifying lazy behaviour because I have acknowledged a fundamental truism about myself: I am lazy at heart. I think it has given me a unique perspective on life. I can spot “lazy” and the subsequent excuses for laziness like Keanu Reeves in The Matrix can see the code. I don’t see dead people, I see the green lazy code.

Navigating life with AI

When I was a kid my father used to give me tasks to do, mostly everyday chores, and if I didn’t do them properly the first time he made me do them again. So I learnt to do things really well the first time, so I never had to do them again.

At that stage of my life, once I’d finished my chores the day was mine. And I mostly spent it reading science fiction books and comics. I never noticed that I applied the same methodology through the rest of my life when it had become a conveyor belt — as soon as you finish one task there is always another task to do.

It turns out my childhood realisation that you have to do the job well has stood me in good stead. As a young reporter I plugged away at each individual task (story). But life is a conveyor belt.

I was forced to coin the phrase “high-functioning, lazy person” to describe myself to my friends. I know the acronym HFLP doesn’t roll off the tongue, but it’s a legitimate personality type. It’s been very helpful to know this about myself.

I learnt when I was young what my fundamental motivations were (to read sci-fi) and how to achieve them (do the tasks my father set); as well as an important corollary (do the tasks well, first time). The latter is obviously key.

Doing something sloppily or haphazardly is a shortcut. Generative AI is a cognitive shortcut. I fear it will only come back to bite the people who use it at the expense of learning the real skills – and therefore acquiring the cognitive benefit – in later life.

I could write a whole book on the tips and tricks in life I have learnt from being a high-functioning lazy person. There are so many ways to get the job done smarter rather than harder. Using GenAI isn’t one of them when it comes at the expense of learning useful life lessons.

As an aside, I found a way to read while doing mundane things such as driving my car and schlepping stuff around my house. Using Audible, I listen to someone reading the book to me, which is the 21st century way of consuming literature.

So, in essence, I’m still getting what I always wanted as a kid — more time to read sci-fi novels.

• Shapshak is editor-in-chief of Stuff.co.za

Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.

Comment icon