As British job seekers turn to AI to help them tailor their CVs and cover letters more quickly, a linguistic expert from dictionary site WordDB has revealed the telltale signs of AI-generated text—and the words you should avoid using if you don’t want to get caught using AI.
Nicoli Westwood from WordDB explains, “While AI platforms like ChatGPT and Microsoft’s Copilot might seem great at generating human-like text, they tend to overuse certain words and phrases.
“If you’re not familiar with AI-generated text, these words might look normal, but they can be a dead giveaway for those who know what to look for. Recruiters can spot these words a mile away and may immediately reject your application if they suspect AI usage.”
If you’re concerned that a recruiter might realise that AI had a hand in your application, then you should avoid using the following words:
- This word is used by AI tools to mean “to explore” or “to investigate.” Its overuse by Large Language Models (LLMs) like ChatGPT makes it a red flag for AI-generated content. For example, “I delved into the client’s needs for this project and developed a comprehensive solution.”
- Similar to “delve,” AI platforms frequently use this word to mean “to encourage” or “to promote.” Some examples might include phrases like “fostering innovation” or “fostering collaboration.” Human writers use this word sparingly, so its appearance in your CV can be a sign of AI-generated content.
- This is another popular choice for AI tools. It’s typically used in more formal content to describe processes that have been optimised for efficiency. AI loves to overuse this word and it can include it in a way that sounds robotic and out of context—which makes sense if the text is written by a robot.
- If your CV or cover letter includes phrases like “beacon of creativity” or “beacon of knowledge,” then recruiters are likely to run a mile. This word is a favourite of AI platforms, but it’s not something human writers use that frequently in conversation or writing.
- AI tools tend to use overly grandiose language, which is why words like “testament” are frequently seen. For example, “This was a testament to my dedication and hard work.” While there’s nothing wrong with this word, its frequency in AI-generated content makes it a red flag for authenticity.
- Underscore is another AI favourite. For example, “My experience in customer service underscores my ability to manage client relationships effectively.” It’s usually used to emphasise a point or idea, but it’s slightly grandiose for everyday writing—something AI doesn’t understand yet.
- Similar to “streamline,” this is another popular word used by AI when it wants to say “to make easier” or “to assist.” It’s not an uncommon word, so it doesn’t necessarily mean it’s written by AI, but it can be a red flag for recruiters.
WordDB’s Nicoli Westwood adds: “These words aren’t inherently bad, but they can signal AI involvement to recruiters who are familiar with these tools. To avoid raising red flags, you need to review your CV and cover letter carefully to make sure the language feels natural and appropriate for the job you’re applying for.”
Beyond word choice, Westwood notes other signs of AI-generated text that job seekers should be aware of: “AI often relies on formulaic structures, such as ‘by doing X, I achieved Y’ or phrases that end with a moral reminder, like ‘it’s important to remember that…’. These can make your application immediately jump out as AI”
Another key giveaway is spelling. “AI tools often default to American spelling, which can be a clear indicator on a British CV. Make sure to double-check that your application uses British spelling conventions to avoid this mistake,” Westwood advises.
So, why does AI tend to overuse certain words and phrases?
“AI models are trained on massive datasets, which include a wide range of texts from books, articles, and websites. These models learn to predict which words are most likely to follow one another based on the patterns found in this data.
“This often means that AI will use predictable words and phrases because they are simply ‘predicting’ the most likely word to follow. This can make AI-generated content quite recognisable for anyone who has worked with these tools.”
