Your AI-Generated Document Could Cost You — Just Ask This Singapore Husband
Imagine submitting a crucial business proposal, only to discover a key statistic was completely fabricated by the AI tool you used. That’s the nightmare a Singapore man just lived through in court — and his lesson applies directly to your business. If you’re using ChatGPT or similar tools to draft contracts, emails, or reports, this story is a wake-up call.
TL;DR: Generative AI can produce convincing but completely false information. A Singapore magistrate just reminded everyone that you — not the algorithm — are responsible for the output. For SME owners, this means every AI-generated piece of material needs human verification before it goes out. No shortcuts.
What Happened in That Singapore Court?
A husband seeking a protection order against his estranged wife claimed she had left him hungry, broke, and physically weakened. But the magistrate wasn’t convinced — partly because the man’s own testimony (carrying 15 kg of flour over 1.2 km) contradicted his claims of brittleness and weakness (source).
What caught the court’s attention beyond the facts of the case was how the husband prepared his legal documents. He had used generative AI to draft his submissions. But the AI invented a legal case that never existed — a classic “hallucination” — and the husband submitted it without catching the error.
Magistrate Soh Kian Peng gave him a public reminder:
“It bears emphasising that litigants are not prohibited from using generative AI. But they must take responsibility for its output, especially where this is evidence that will be presented to the court. The very issue with using generative AI in the drafting of such documents is the temptation to rely entirely on the algorithm without applying one’s mind to the actual output.” (source)
The court placed no reliance on his AI-assisted statements. The man lost his case and was ordered to pay costs.
The Real AI Danger for Business Owners
This isn’t a one-off oddity. Large language models like the ones behind ChatGPT, Claude, and Gemini are designed to generate plausible text — not verified facts. They frequently:
- Cite nonexistent cases, studies, or statistics.
- Invent quotes and attribute them to real people.
- Create convincing but false timelines or figures.
If you’re a business owner using AI to draft client proposals, service agreements, marketing copy, or even internal policies, you’re exposed to the same risk. A hallucinated number in a financial report or a fake testimonial on your website could damage your reputation or land you in hot water. The Singapore husband’s mistake is a mirror for what could happen in any SME.
Why This Hits Home for Malaysian SMEs
You and your small team probably wear many hats. It’s tempting to ask an AI to “write a professional email”, “draft a refund policy”, or “create a social media post”. But here’s the rub: you are still legally and professionally responsible for every word that leaves your business.
Consider these scenarios where AI hallucinations could hurt you:
| Scenario | What the AI might invent | Who gets blamed |
|---|---|---|
| Client contract clause | A compliance requirement that doesn’t exist | Your business |
| Product description | An award or certification the product never earned | Your brand |
| HR policy manual | A legal reference that is outdated or fake | You (as employer) |
| Customer support email | A fabricated refund timeline or technical detail | Your team |
The pattern is always the same: the AI produces something that looks right, but the business owner fails to verify it. That’s the recipe for a disaster that feels like being publicly called out in court — or worse, by your own customers.
How to Use AI Without Getting Burned
The magistrate in Singapore didn’t ban AI in court. He just demanded accountability. You can take the same approach in your business. Here’s a practical checklist to make AI your assistant, not your liability:
- Always verify specific claims. If the AI mentions a regulation, a statistic, or a legal case, go find the primary source before you use it.
- Read the whole document through. The Singapore husband’s documents switched between “I” and “he”. Odd inconsistencies like that should trigger a full review.
- Ask the AI for its sources. Many tools now provide citations or links. Click them. If it can’t provide a real source, treat the information as unverified.
- Keep a human in the loop. Use AI for first drafts, not final versions. Someone with domain knowledge must sign off on anything that goes to a client, partner, or authority.
- Build verification into your workflow. If you regularly generate certain types of content (e.g., quotes, proposals), create a checklist that you or your team runs through before publishing.
These steps don’t slow you down much — they protect you from the kind of embarrassment the Singapore husband faced.
The Bigger Picture
This court case is a microcosm of a much larger truth: AI is a powerful accelerator, not a substitute for judgement. As tools get better at sounding human, the temptation to “trust the AI” grows. But every business owner knows that trust must be earned — and verified.
The same principle applies whether you use AI for court documents, customer emails, or internal memos. The technology is here to stay, but so is accountability. The skill that will separate thriving SMEs from the ones that slip up is not how fast you can generate content — it’s how carefully you review it.
We help SMEs find the right balance between automation and human oversight. If you’re building AI-powered processes for your business and want to make sure they’re safe, effective, and built for real-world results, let’s talk.
Book a free 15-min call to see how [topic] applies to your business →