A new kind of practitioner has emerged in the age of generative AI. They prompt, copy, paste, and publish. No verification. No critical thinking. No human layer added anywhere in the process. They function almost like conduits, not contributors, not value-adders.
We have an exact parallel for this right here in Singapore.
Walk into a Taoist temple during one of the religious festivals and you might catch a spirit medium at work. These are people who enter trance states to channel deities, granting blessings or dispensing lucky 4D numbers to whoever is waiting.
So what exactly gets lost in transmission?
Fact-checking, for one. AI hallucinates with complete confidence. If you don’t verify before you publish, you’re the one who looks foolish, not the model.
Then there’s contextualization. AI knows about Singapore’s hawker culture. It has never waited forty-five minutes for Tian Tian chicken rice on a Sunday. There’s a difference.
Critical analysis is another gap. AI summarizes well. But summarizing and thinking are not the same thing.
Ethical filtering too. AI will give you options. You’re the one who has to live with the choice.
Cultural fluency matters more than people realize. Ask AI to clean up a design and it might suggest removing red to reduce visual noise. It won’t know that red means luck and prosperity for your Singaporean Chinese client, or that stripping it from a CNY campaign without a conversation could quietly sink the whole thing.
Tonal nuance is just as invisible until it isn’t. The loaded shrug in a “can, lor.” The passive-aggressive warmth of certain corporate Singaporean emails. AI misses these constantly. A perfectly grammatical apology that reads like a legal notice is still a bad apology.
Moral trade-offs don’t resolve themselves either. When profit and people are in tension, AI gives you a balanced list. Someone still has to make the call and carry the weight of it.
And experiential memory. AI has data. It doesn’t have scar tissue. A veteran marketer who remembers a similar campaign flopping in 2019 catches things no model will ever flag.
Instead of channeling AI output, it helps to be deliberate about the role you’re playing. Be the editor who curates, cuts, and sharpens what the model produces. Be the verifier who checks the facts, the sources, and the logic. Be the synthesizer who combines AI output with your own knowledge and judgment. Be the ethical guardian who applies the moral compass the model does not have.
So when you use AI, what role are you actually playing? Because here’s the uncomfortable question: if you removed yourself entirely from the process, would anyone notice? Would the output be any worse? If the answer is no, you’re not a professional using a tool. You’re a middleman who learned to type prompts. The real work isn’t in the prompt. It’s in everything you bring between the prompt and the publish button. Because at the end of the day, ask yourself, whose name is printed at the top?
