A cautionary tale emerged in August 2025 when Rishi Nathwani KC, a senior barrister in Victoria, Australia, was publicly reprimanded for submitting AI-generated legal arguments in a murder trial. The submissions included fabricated quotes from legislative speeches and fictitious case law, purportedly from the Supreme Court. Justice James Elliott delayed the case by 24 hours after court associates failed to locate the cited precedents. Nathwani admitted the citations “do not exist,” having assumed their accuracy based on a few verified entries. The fallout was swift and sobering: a reminder that even seasoned professionals can be misled by AI’s confident tone and polished output.
Artificial intelligence has rapidly become a staple in the business copywriter’s toolkit, offering speed, scalability, and a surprising knack for tone-matching. When used wisely, AI excels at drafting articles where the human author provides the facts, structure, and intent—allowing the machine to handle the linguistic heavy lifting. This is especially effective for internal communications, product descriptions, and marketing blogs where the subject matter is well-understood and the factual base is solid. In these cases, AI acts as a tireless assistant, rephrasing, summarizing, and formatting content with impressive efficiency.
But the reliability of AI in copywriting hinges on one critical factor: the truth must come from the human. When tasked with generating non-fiction content independently—especially in technical, legal, or historical domains—AI can veer into dangerous territory. It may fabricate plausible-sounding details, statistics, or even citations. These so-called “ghost citations” are references to sources that don’t exist, often presented with convincing formatting and tone. The risk isn’t just academic—it’s reputational.
This incident underscores a broader truth: AI is not a source of knowledge, but a pattern generator. It doesn’t “know” facts—it predicts what words are likely to follow based on its training data. In business copywriting, this means AI should be used to express what the writer already knows, not to discover or assert new truths. The best practice is to treat AI like a junior editor: helpful with phrasing, formatting, and tone—but never trusted to originate facts or verify sources.
As AI tools become more embedded in business workflows, the burden of truth remains firmly on the human. The pen may be digital, but the responsibility is not.
Leave a Reply