‘Hey ChatGPT, is Gen Z the canary in the AI coalmine?’

Artificial Intelligence models that create words, images and videos may have just become available across most devices and platforms, but they are already a hit with users. And with the technology yet to be properly regulated, and its long-term effects yet to be fully understood, this world of early adopters, while exciting, is an equally unsettling one.

From bots that write emails and summarise messages for you, to image-generation apps that can create selfies in locations you’ve never visited, the AI craze is moving fast and appeals particularly to young people who feel they need to prepare for an AI future.

Research published in October based on a survey of university students suggested that those in Gen Z were significantly more likely to use ChatGPT (for example, to help draft or structure papers) during their courses, relative to other generations. Studies have also been conducted to learn more about how the technology affects learning comprehension and critical thinking, though a consensus is far from being reached.

AI automations are all the rage, but they can be a lot of work for very little result.Credit: Pexels

One study from this year (notably conducted by a company that provides AI solutions) suggested four out of five Gen Z workers used generative AI at work, and a quarter of Gen Z workers delegate half of their workload to AI, though anecdotally this would depend largely on the kind of work being done.

Young enough to have grown up with ubiquitous internet, but old enough to be entering the workforce, Gen Z could ultimately be seen as the canary in the coalmine when it comes to generative AI. But how will they choose to use it?

Dr Dana McKay, senior lecturer in innovative interactive technologies at RMIT University, said it might not ultimately be up to them.

“Companies are pushing [generative AI features] at us. Young people, especially in the tech space, are quite open to using them. And they do have benefits for some people. But should we be using them for everything?

“Tech companies are sticking them in all sorts of places. In our email, our word processors, our search results. It’s not whether Gen Z is choosing to use it, it’s whether they’re going to have the choice not to use it.”

Google’s search engine now includes many generative AI features, which have been panned as unreliable but which may use up to 10 times as much energy as regular searches.Credit: Google

Quick answers and quick money, but with caveats

New AI features continue to appear every day, despite the potential of a dire environmental cost. When Google rolled out its AI-powered summaries in Search last year, reports indicated the energy requirement of a simple search increased by 10 times.

But for those engaged in the space, Google may be becoming a thing of the past anyway. For the simple kind of information-gathering query we’re used to dropping into a search bar to find an answer on Wikipedia or Reddit, ChatGPT and similar bots have become a popular alternative.

Asked what sci-fi movies are now showing, Chat-GPT very confidently gives a completely incorrect response. In this case, Google is better, even though it also didn’t know; it linked me to cinema websites.Credit:

A cottage industry has sprung up around teaching people how to optimise their queries to get the information they need, and in theory AI can scrape numerous sources to come up with a consensus rather than the user manually having to click multiple links. But chatbots also frequently get facts entirely wrong, despite being designed to appear objective and completely confident.

“Being used to something producing an answer that sounds certain probably will change expectations of what an answer looks like,” McKay said, potentially making it difficult to go back.

“There are things in life that are genuinely disputed, and how we ask an AI to address something that’s genuinely disputed is a real challenge in that environment.”

The ubiquity of generative AI has also given rise to an ecosystem of AI influencers, showing their followers how to make money with AI, kickstart their hustles and make their lives more efficient, either with original content or by highlighting useful titbits shared elsewhere on the web.

Unsurprisingly, a lot of the most popular posts are get-rich-quick schemes or ideas for passive income, which is a big reason many online content platforms have begun to fill with AI slop. A recurring example suggests that people ask ChatGPT to make a script for a podcast based on today’s news, then plug that script into a voice generator. Add automation software that can repeat the same tasks on a set schedule and upload the results to a podcast service, and you have a low-quality regular podcast that is very unlikely to attract listeners or ad dollars.

On the other hand, there’s a lot of more constructive AI influencer content too.

A technology communicator who goes by the name AskCatGPT discusses practical tips for using the technology, but also cautions against potential negative implications. In one video, for example, she uses an automation app called Zapier to have ChatGPT draft replies to every email she receives, while in another video she relays conversations with young teens who worry that their AI use is becoming a crutch that will stunt their critical thinking skills.

AI in the classroom

Of course, tech influencers on TikTok may no more represent Gen Z than Mark Zuckerberg represents Millennials, and there’s reason to think teens may tire of many AI trends before they’re even eligible to vote.

Nick Donaldson, a high school English teacher from Melbourne’s east, said his students had already been through several cycles of AI adoption and use.

“When chatbots were first introduced, the most common use seemed to be what we were expecting: a few students using the tech to generate entire pieces of work. We used apps to identify AI-generated content and actively discussed chatbot use with the students, highlighting the consequences and exploring possible appropriate uses,” he said.

“Now they seem to be using chatbots in increasingly peripheral ways. Instead of getting them to create work whole cloth, they might ask a bot to share initial ideas or mark work based on criteria, or offer guidance on how to improve based on the provided task requirements.”

Donaldson said AI has a place in the classroom, and he himself uses it with appropriate care when preparing some educational materials. But while he’s seen many students using AI in place of search engines or spell-check, he said they would benefit from education in how to craft effective prompts, how to use AI for study without it becoming plagiarism, and how to fact-check AI or spot false information.

On the other hand, some students have found the pervasiveness of AI tools overwhelming, and prefer to avoid them for certain tasks.

“In the last year or so, I’ve noticed an increased request to handwrite work instead of crafting it digitally. Students seem keen to take a break from not only AI tools but computer technology in general,” Donaldson said.

“It’s a similar case with mobile phones. Recent bans in schools have been pretty positively received by the student community. So yeah, they seem to see the value in AI and digital technology, and use it when it suits, but they also seem to be increasingly seeking solace through more hands-on, tech-free approaches.”

Where are we headed?

What about the idea that generative AI could be used to augment interpersonal communication? Already there are services that will reply to emails for you, summarise incoming messages, or answer the phone and transcribe what’s being said. There are even generative AI tools for crafting dating app profiles and messaging with prospective matches, which would have wild implications if everybody was using it.

It’s not uncommon to hear from older generations that Gen Z is socially awkward, withdrawn, hates talking on the phone or face-to-face and so on. So, isn’t this the perfect match?

Clearly there are certain tasks people will prefer to use AI for, just as many people have now moved to Googling symptoms rather than talking to a doctor, or asking Reddit rather than asking real-life friends. And for this kind of use case, the technology would seem to have unique benefits as well as dangers.

Research published in Science in September found that chatbots specially trained to debunk conspiracy theories had a phenomenally high success rate at changing the opinions of people who believed them. The AI is patient, persistent, consistent, and was seen by the conspiracy believer as dispassionate and informed. This is a good sign that conversations with AI could actually be preferable to social media bubbles in some instances. But it also shows that, depending on who designs the AI system, it could be used to reinforce dangerous beliefs.

McKay, the RMIT lecturer, said it might not be entirely useful or accurate to say younger generations are “less social” than previous ones, and therefore more likely to turn to AI. Rather, younger generations have grown up with more options and tools to get things done and connect with others.

“What we grow up with changes our expectation of what’s possible in the world. Gen Z grew up in a search era. And any kid under the age of 10 has been searching since before they could read or type, by mashing the buttons and asking the questions that their parents said they didn’t have time to answer right now,” McKay said.

“Some of this is about different input modalities. A lot of this is about support when writing on a small screen. It’s horrendous to type an email on a piece of glass that’s just over the size of my hand. And so the generative AI helps me type that email, but it also makes me sound less like myself.”

Will AI eventually fulfil that great technological promise of letting humans get on with the human things while machines handle the tasks that are easiest to automate? It may be too early to tell. But tapping into what the early adopters and the influencers are using it for, an awful lot of it seems to involve the opposite: people working hard to coax machines into creating art and being personable.

McKay was hopeful that as the technology matures it would find the spots in each industry where it could do the most good. But she also suspects that AI, like email, may create as much work as it alleviates.

“Say you’re using it to summarise something, but you’re doing that in academic work. Well, then you have to fact-check the summarisation, right? So do you actually save any time, or do you spend the time fact-checking instead of just reading the damn thing and developing your own opinion in the first place?

“There’s a lot of talk about how we measure and educate people’s critical thinking around what AI produces, which, again, is new work that we’re going to have to do. So right now, it feels like it’s going to be this massive labour-saving device, but I think it’s just going to change the nature of a labour.”

Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.