.jpg)
By Mihar Dias (C) September 2024
In the rapidly changing world of artificial intelligence, mistakes are inevitable. Teresa Kok’s recent experience with ChatGPT, where an error in translation from Mandarin to Malay sparked controversy over the halal status of restaurants is a sobering reminder that while AI tools like ChatGPT can be powerful allies, they are far from infallible.
It also raises important questions about the role of human oversight in AI-generated content and the potential consequences when that oversight lapses.
Kok’s explanation that her Malay staff, pressed for time, resorted to using ChatGPT for translation is understandable. AI tools like ChatGPT offer a quick and easy way to handle complex linguistic tasks, especially when deadlines loom.
However, as Kok herself acknowledged, the failure to thoroughly review the AI-generated text led to the public relations debacle. The mistranslation of key phrases not only misrepresented her original statement but also allowed the media to seize upon the error, amplifying the controversy.
This incident is not an isolated case. As AI becomes more integrated into professional and personal work, particularly in areas like translation, communication, and content creation, the risks of errors like this increase.
Language is subtly linked with cultural and contextual factors that even the most advanced AI models struggle to fully grasp. When humans delegate these tasks to machines without adequate supervision, the potential for miscommunication—sometimes with serious consequences—becomes all too real.
The problem is not with ChatGPT itself. AI models, by their nature, learn from vast datasets and are designed to handle language with incredible speed and efficiency.
Yet, as sophisticated as they may seem, they are tools, not replacements for human judgment. ChatGPT can generate text in dozens of languages, but its understanding is based on patterns in the data it’s been trained on, not a true comprehension of meaning.
As a result, subtle differences in tone, context, and intention can easily get lost, especially when translating between languages as structurally different as Mandarin and Malay.
Kok’s situation underscores the importance of recognising AI’s limitations. In this case, the technology was treated as a shortcut—something to fill in when human hands were unavailable. This approach, while tempting, is fraught with risks, particularly when handling sensitive topics like halal certification, which touches on deeply rooted cultural and religious concerns. A mistranslation, even if unintentional, can be seen as disrespect or incompetence, fueling misunderstandings that can spiral out of control.
The lesson here is clear: AI tools like ChatGPT are best used as assistants, not autonomous agents. Human oversight is not just advisable but essential, especially when dealing with matters that require precise communication.
A tool like ChatGPT can offer speed, but accuracy still requires the discernment that only humans can provide. Kok’s statement, after all, should have been reviewed before it went public—not just for grammatical accuracy but for its tone and messaging.
In the broader context, Kok’s experience also serves as a reminder for policymakers, businesses, and public figures: relying too heavily on AI without ensuring quality control can backfire.
As AI tools become more commonplace, it is crucial to establish best practices for their use. This includes reviewing AI-generated content, understanding the risks involved, and ensuring that sensitive or high-stakes communications are double-checked by human editors.
Ultimately, the ChatGPT mistranslation in Teresa Kok’s case wasn’t the root cause of the controversy—it was the human decision to rely on AI without sufficient oversight. And in an age where AI is increasingly at the forefront of public life, it’s a mistake that all of us should learn from.
Mihar Dias is a content creator under the Newswav Creator programme, where you get to express yourself, be a citizen journalist, and at the same time monetize your content & reach millions of users on Newswav. Log in to creator.newswav.com and become a Newswav Creator now!
The User Content (as defined on Newswav Terms of Use) above including the views expressed and media (pictures, videos, citations etc) were submitted & posted by the author. Newswav is solely an aggregation platform that hosts the User Content. If you have any questions about the content, copyright or other issues of the work, please contact Newswav.
