By: Yoyo Rita
AI usage is rising quickly, and profoundly, in the workplace. Most of us can assume this as a common-sense fact—after all, many of us use AI tools ourselves in our jobs. Regardless, it’s helpful to understand just how profound a shift in AI usage is occurring in workplaces across all industries. A recent Gallup study[1] reveals as much, showing a nearly 20% increase in infrequent workplace AI usage from 2023-2025, and a nearly 10% increase in more frequent workplace AI usage from 2023-2025:
This uptick in both regular and occasional AI usage in the workplace is especially true for white-collar workers over the past few years. The data shows that white-collar workers’ frequent AI usage went up 15% just in the past two years. Take a look at the Gallup data:
Ok, that’s interesting (albeit predictable) enough, considering AI’s advanced administrative functions that assist greatly in white collar contexts—but what changes can we expect to see in workplace investigations?
As many investigators have already begun to notice, we can expect the greater adoption of AI in workplaces to result in more AI-written workplace complaints. Indeed, Complainants are increasingly relying on AI assistance to streamline and refine their complaints.
As investigators, we must develop a trained, discerning eye to better understand how AI-generated complaints and other documentation should be handled in our processes.
Here are a few recommendations and rules of the road to spot such complaints, and to know how to handle them.
- AI-written and -assisted workplace complaints are usually easy to spot. They invariably include bolded categories delineating major categories like the timeline of alleged events, the context, the impact, and the possible policy violations therein. Their syntax tends to be didactic, clinical, and avoiding anecdotal asides. Because AI text detectors tend to be unreliable, use your best judgement (often based off of diction, structure, and repetitive formatting) when determining if a complaint was likely written by AI.
- Whether a complaint was written by AI or not, investigators must treat all complaints with the same standards of impartiality, thoroughness, and prompt fact-finding. Our standards of practice do not change, regardless of the new technological environment we are immersed in. However, the complexity of the complaints that AI tends to generate means an investigator will have more data points to cover from an AI-drafted complaint, as compared to your standard human-drafted complaint.
- When you suspect a complaint was written by AI, ensure you check each and every data point in the complaint. This goes for interviewing the Complainant and other witnesses. AI hallucinations, data errors, and extrapolations can result in a complaint that the Complainant themselves may not be aware of, or even in agreement with.
- Keep track of any discrepancies between the written complaint and the Complainant’s testimony. If needed, an investigator should give the Complainant the opportunity to clarify the specifics of their complaint if AI misconstrued any aspect of their claims. However, in such a case where a change would result in a marked material difference in the allegations, this would require a thoughtful approach by the investigator—including, in some cases, directly asking the Complainant if their complaint was drafted by AI.
- Trust your instincts and experience. Our existing ethical codes and standards of practice as investigators provide a valuable framework to assess credibility and to investigate the likely veracity of a set of facts. This framework will serve investigators well in the increasingly complex landscape of AI-supported workplaces.
All Complainants, Respondents, and parties to an investigation are entitled to a fair, impartial, and thorough investigation process. While AI usage may complicate an investigation in some cases, it is our responsibility to act ethically and effectively no matter the facts, or the manner in which they are presented.
[1] https://www.gallup.com/workplace/651203/workplace-answering-big-questions.aspx