A Canadian lithium exploration company’s recent corporate governance statement, intended for its debut on the Australian stock exchange, drew unexpected attention last month after containing a garbled sentence that appeared to be the result of faulty artificial intelligence (AI) use. The company, Li-FT Power, did not respond to inquiries regarding the error, but the incident highlights the ongoing challenges businesses face when integrating AI technologies not originally designed for corporate applications.
Throughout 2024, AI-related errors and controversies have become increasingly common across various sectors, from politics to publishing. Public figures have frequently been scrutinized for alleged reliance on AI in their work. Former Reform UK candidate Matt Goodwin, for example, denied claims that AI contributed to his immigration book, yet the episode earned him the nickname “MattGPT.” Similarly, Labour MP Mike Reader has been dubbed “ChatG-MP” following reports that he used AI tools on a train to reply to constituents. In the United States, publishers canceled the release of the novel *Shy Girl* amid suspicions that AI had been used to assist its writing.
Journalistic fields have not been immune to fallout, with multiple reporters losing jobs or facing suspensions over undisclosed AI use earlier this year. Meanwhile, US media entrepreneur Steven Rosenbaum disclosed that AI generated fabricated or misattributed quotes in his book *The Future of Truth: How AI Reshapes Reality*, underscoring the technology’s potential to distort information even in professional contexts.
Perhaps most consequentially, the legal sector has seen growing challenges stemming from AI errors. In March, a family dispute over an Oregon winery led to lawyers being fined $110,000 after citing fictitious cases produced by AI tools. Magistrate Judge Mark D. Clarke described the case as an extreme example in an evolving landscape of sanctions for AI misuse in legal proceedings. Damien Charlotin, a French lawyer and academic who maintains a database tracking AI-related legal missteps since 2023, confirmed that the Oregon penalties are among the highest recorded. His database has expanded from fewer than 700 cases at the end of 2023 to over 1,500 as of June.
The surge in AI-related errors coincides with an increase in litigants representing themselves in US federal courts, rising from 11 percent to 17 percent, which experts warn may contribute to backlog growth. Charlotin acknowledges the utility of AI in automating routine tasks, allowing professionals to focus on more complex, human-centered work. However, he emphasizes that large language models inherently "hallucinate" or generate false information, requiring careful human oversight and narrow task-specific applications.
As AI becomes more widely adopted, its improper use and overreliance pose ongoing risks across industries. Experts caution that despite its benefits, AI must be managed responsibly, with humans remaining firmly in control. The persistence of missteps like those seen in governance statements, political communication, publishing, journalism, and legal proceedings demonstrate that the technology’s growing pains are far from over.