One Prok Chop Cheese Baked Rice, please!
“Cheese Baked Rice (Prok Chop/Chicken Chop)” was printed on the menu of a Hong Kong-style restaurant. Two years ago, I would scowl at the spelling mishap and roll my eyes at the thought of a copywriter making such a mistake. Instead, I stared in amusement at the laminated booklet. “How endearing. They didn’t use AI for this,” was my immediate thought as I flipped through the pages, each containing photos of food that did not appear to be lit the same way. There was an undeniable charm in knowing the menu was not AI-generated. Peep the uncanny illustrations where the texture of the burger bun blends into the patty, or how certain products do not look as advertised. I once copywrote for a now-defunct club, where I churned out lineup announcements and DJ introductions. Its formulaic approach to socials, which I secretly relished, was perhaps also the reason that engagement ended, as I learnt it was significantly more cost- effective for the team to generate captions than to keep a writer on retainer. Beyond just commercial and creative spaces, Generative AI (GenAI) has proliferated in our daily lives — more than 60 per cent of Singaporeans use the technology, and Malaysia is the top Southeast Asian country for generating images with AI. The sentiments towards the technology are polarising — tastemakers express their disdain, while business owners salivate at the thought of increased productivity at presumably lower costs — but how true are these narratives? Men’s Folio engages with policy researchers across Singapore and Malaysia to learn and understand more.
View this post on Instagram
Creating AI fruit slop, race-swapping pop stars, and using “Chat” for trauma dumping are all capabilities of GenAI, a facet of AI that creates text, photo, and video responses based on user prompts. Claude and ChatGPT are AI tools powered by Large Language Models (LLMs), enabling users to generate content and predict text from massive datasets. They sound human because, well, they are trained to replicate humans. “From being a text-only platform in 2022, ChatGPT has gone on to create images, videos, websites, and even deploy agents that can access other applications to manage your life or business,” explains Wei (pseudonym), a Singapore- based policy analyst at a regional nonprofit. “There are also other more specialised, domain-specific models for, say, healthcare, law, and a growing field of physical AI, which integrates the technology with the physical world through robotics, such as autonomous vehicles, drones and Industry 4.0.” Beyond personal uses of research and writing, the professional services industry uses AI for customer service and sifting through internal records.
The undeniable concern many have about GenAI technologies is that they are displacing human labour. Compared to the past, when innovations like the introduction of typewriters and digital cameras rendered skills like typesetting and photo developing jobs nearly obsolete, how is the widespread introduction of GenAI any different or similar? Is it also a “necessary evil” that will eventually make us all better off? “Every major technological wave has winners and losers. Some jobs disappear, others transform, and entirely new occupations emerge,” says Hanson Chong, an economic policy researcher at the Institute of Strategic & International Studies (ISIS) Malaysia. Given the greater exposure to LLMs among clerical workers and professionals, he sees LLMs’ ability to automate routine manual work and perform cognitive tasks as a positive development in principle, as it gives us more time to spend on things that matter. Dr. Jun-E Tan, senior research associate at Khazanah Research Institute (KRI), proposes a broader view. “LLMs are interesting because they’re general-purpose, compared to the specific functionality of the aforementioned innovations, and are impacting more than just work alone — it seeps into culture and the way we communicate. If we look at society as a string of connected individuals, this particular technology impacts the network between these individuals,” she explains. Notable examples of AI breaking connections include social media algorithms that keep us within certain echo chambers, or jobs like rideshare drivers, who are considered partners rather than employees with a line of reporting. “They’ve only got a screen as an interface and have lost a formal connection to their peers and other workers, nor do they need a connection with their customers.”
The question of whether GenAI’s output is even good remains. Is it actually creative? Are you proud to serve it? If the whole basis of LLMs is to aggregate vast amounts of data and then predict based on statistical likelihood, does that not mean they can only produce something… mid? Our panels of experts agreed unanimously. “GenAI is meant to produce averages based on what they are trained on as ‘good writing’. There are certain things deep in their training data that they can’t get rid of, like em dashes. This brings up a bigger philosophical question on whether creators now need to do something else to distinguish their output from AI, but setting these boundaries is going to be difficult,” Wei expresses. “Before AI, if you created something really unique and special, others could copy your style. But now with AI, they can just do it much quicker than before, without any real skills. At least before, forgers also needed to be painters.” Chong sees LLMs as effective only for producing conventional answers supported by existing information and less reliable at producing breakthroughs. Given the lowering cost of producing content, Dr. Tan observes a trend towards the enshittification of content where everything just becomes slop. It always comes as a surprise to me when people are gratified by the output, so much so that they consider it an unprecedented discovery. When it comes to the arts, how have we become so quick to place our trust when creativity has always risen from the tensions of the lived human experience — just as I have written from a place of being inspired by personal heartbreak, the joys of being in a community, or the wit of Wendy Williams? For corporations that have been quick to replace their writers, designers, photographers, and videographers with AI technology, are they driven purely by cost, greed, or other beliefs?
View this post on Instagram
If one considers conversations about the subjectivity of creativity moot, we can instead discuss the factuality of GenAI responses. “It’s easy to fall into the trap of thinking LLM-generated responses are accurate, because the responses sound fluent and confident. It’s also easy to assume that the responses are objective or neutral, when these outputs are shaped by training data design choices and the prompts given,” shares Chong. The pseudo-confidence exhibited by LLMs is dangerous because it used to be obvious when someone did not know what they were talking about. “There would be thin research, holes in their words, and leaps of logic. Our whole lives we’ve developed a heuristic of judging whether something is correct or not based on how it sounds,” Wei states. “If the AI model is correct 99 per cent of the time, you might miss the one per cent where it’s not correct — especially in areas you don’t have the expertise in, which is where most of us are using AI.” Given that models are mostly trained on Western material, you might not even achieve meaningful findings on specific domains. “We also know that AI reflects and magnifies existing biases, from women to minorities, and so much more.”
Given these considerations, does GenAI still posit itself as a time- and cost-effective alternative to a human workforce? Chong believes it comes down to the tasks. “There are tasks that LLMS can outperform humans on, such as processing large amounts of information, identifying patterns in them, and handling large repetitive tasks at a scale that would be impossible for an individual. However, they aren’t better at every kind of work, especially ones that require judgment and decision-making — they also cannot be held accountable for consequences. In this case, human-in-the-loop remains important,” he shares. If the purpose of adopting LLMs is to optimise, then Dr. Tan implores others to really consider why, and what we are optimising. “There have been cases of companies firing humans and rehiring them back eventually, when they realised that it was not realistic. They may be slower, but speed might not be the only goal. If it’s a customer service job, you want a human there to demonstrate empathy and actual understanding — we all know how infuriating it is to speak to an AI bot,” she shares. Cost- effectiveness is not a guarantee either, as there have been instances where big tech has pulled back on AI adoption due to token and other associated costs exceeding hiring costs.
In Brett Hemenway Falk and Gerry Tsoukalas’s paper, The AI Layoff Trap, they discussed whether companies would suffer a drop in demand stemming from reduced spending power due to unemployment if AI were to displace the human workforce. They also hypothesised that companies will not slow down on AI automation, as it is an arms race that no market force can break. While Chong does not foresee a broad recession driven by mass unemployment in the short term, he does not rule out that possibility in the longer term. “If AI adoption primarily benefits owners of capital and technology, then a larger share of income may flow towards those groups rather than workers. Whether this leads to fewer jobs depends on how the economy adjusts: if AI enables the creation of new products with lower costs, then AI could create new jobs, but if AI mainly produces existing goods and services in markets where demand is already limited, firms may be able to produce more with fewer workers.” What he is curious about is the kinds of new demand AI will create. “Capitalism creates an environment where AI companies have more control over this proliferation in technology than governments and the people’s representatives, so the development of AI is profoundly undemocratic and elitist. Tech and other companies are investing heavily in this, so if that doesn’t pan out, it could create a recession,” shares Wei. Dr. Tan, however, warns against viewing the situation in an oversimplified manner. “We have climate change, wars, and an already bad economic situation, so it’s hard just to pin it on AI. When people lose jobs, you face a situation where they’re trying to be employed for roles they are overqualified for and this disproportionately affects younger people. It’s always about how unequal technology renders things.”
Have we been changed, for good? From being wary of em dashes to calling out brands and creators who use AI, it is safe to say that we, too, have actually altered our behaviours and belief systems with the proliferation of GenAI. “AI has changed how people express themselves, and our expectations of how others communicate with us. I can see AI’s influence through certain patterns, such as the overuse of ‘not only… but also’ sentence structures, or three-part lists like ‘address X, enhance Y, and deliver Z,’” observes Chong. Dr. Tan sees AI technology acting more like mirrors than windows for users, isolating us instead of connecting us. “Art can be windows of lived experience, where we get to access and immerse ourselves in different stories and lives, like when we read a good novel. But when we are fed content that are tailored to our taste and fitted to our personalities, we are surrounding ourselves with mirrors rather than windows”. If anything, GenAI has made us a lot less patient and a lot more narcissistic, as we are now accustomed to the agreeable nature of these models that deliver desired outputs within seconds.
Even if the benefits of AI technology will eventually outweigh its cons and produce a net good in the long run, it does not mean that everyone will benefit equally. “As technology usually comes with a certain power structure, those who can afford the technology will wield it and reap the benefits the most, while the rest of us play catch-up. It can feel like there’s no option to opt out of this race,” shares Dr.Tan Subsequently, Chong notes that AI exposure is lower at higher rungs of the wage ladder, as highly paid professionals are more likely to perform tasks that rely on complex problem-solving, socio-emotional skills, and human judgment. This aligns with Wei’s findings, which showed that individuals with less exposure to technology and education have been using GenAI to create market strategies, visuals and assets such as menus for their businesses. “These are average outputs that won’t win awards but do help with learning and growing businesses. Ultimately, business owners still need some ability to judge the quality of the output to create something meaningful,” notes Wei. Tan sees the possibility of a bifurcated society, where individuals of high socioeconomic status will enjoy analogue craftsmanship, while the rest of us can only afford to experience sterile slop.
To mindfully and meaningfully adopt GenAI is to look beyond what the technology can achieve and ask how it affects people. “Ultimately, it depends on how we choose to adopt and govern it, as AI could deepen inequalities, displace workers without adequate support, and create privacy risks and greater strain on the environment,” notes Chong. While he does not see AI and human creativity as disparate, he recognises that it will take time for society to settle on what role AI should play in art. “Fully autonomous AI-generated content can feel like ‘slop’ because art resonates with us through the human experience, like how we perceive the world, form intentions and derive meaning — processes that AI cannot fully replicate.” While AI has empowered those who previously lacked the resources to realise their dreams, Wei remains uncertain that the benefits offset the impact on people’s livelihoods as they develop their skills.
A now-endearing typo as seen on a local restaurant menu.
Everyone feels the helplessness of potentially being displaced by AI, but I doubt anyone truly believes that rationing their individual use will make a meaningful impact. So where do we go from here? “Individual behavioural change, such as reducing plastic straws to save the environment, will only get you so far. Meaningful changes need a system to support them,” explains Dr. Tan. Countries are quick to focus on productivity and economic gains but still make little progress in managing the risks of AI adoption for workers, communities, or the environment. To mitigate, humans should remain central in decision-making rather than be seen as a cost to be reduced. “AI can provide information and support, but should not replace meaningful conversations with family and friends,” Chong emphasises. Wei recommends collectivising and voicing out concerns, as governments are listening. “The backlash forces companies to be more transparent and to meet higher standards. Owners of capital aren’t going to do the more expensive, responsible thing unless they have no other choice,” offers Wei. In fact, she encourages individuals to reclaim their power. “We have to be smart and realistic with AI. It’s changing our world, but not just as a technology that’s inevitable. For it to scale, it requires people to agree to it and use it, and for company owners to adopt and pay for it. There are many areas we could work on to make a world with AI a better one than the default dystopian scenarios we’re confronted with.”
View this post on Instagram
In a societal landscape that already feels so isolating for so many, it is no surprise that GenAI tools appear like an intoxicating reprieve. It feeds into our desires for immediacy in a culture that forces us to believe we are running out of time; it allows us to affirm our beliefs from a convincing voice that agrees with us for the most part, which feels deeply gratifying. The yardsticks keep moving, so we have traded breakthroughs for a familiar, unimaginative taste that works. Changes feel extra terrifying given how they evolve faster than we can understand them and create proper safeguards. In fact, in the process of writing this, I have come to realise that my biggest fear boils down to humanity being displaced in favour of achieving cost-efficiency for a select few. AI do, or AI don’t? The technology is here to stay, but are you able to see it for what it is?
Once you are done with this story, click here to catch up with our September 2026 issue.