Companies are developing sophisticated systems to identify artificially generated emails as more than 6.7 billion AI created messages are sent daily, creating challenges for businesses trying to manage legitimate communications. Experts have identified key linguistic patterns that distinguish machine written text from human correspondence.
Vahan Poghosyan, CEO and Co-Founder of AI link building software company Linkee, said around 70 percent of partnership texts received in the last year were written by artificial intelligence. To address the volume, his company developed a detection system that saves 1 to 1.5 hours daily on email communication.
“Robot-written emails are sent more and more,” Poghosyan stated. “To deal with so many emails like that, we developed a system that spots generated texts, so we don’t spend as much effort and time on them.”
Experts have identified specific phrasings that signal automated writing. Overly formal greetings are used 42 percent more often by automated writing tools compared to human authors. Traditional email openers like “Dear Sir or Madam” and “I hope this message finds you well” have become signatures of machine written correspondence.
Language models replicate polite formalities effectively but cannot fake genuine sincerity, which matters most in modern email communication, according to experts. They recommend using actual names and getting directly to the point rather than including formulaic pleasantries that signal automated generation.
Artificial intelligence also tends to overuse transition words, creating another detection marker. Even in human written text, loading messages with “moreover,” “furthermore,” and “additionally” makes content sound robotic. These connector words are supposed to enhance clarity, but AI systems use them so frequently they have become red flags for automated content.
Maria Harutyunyan, co-founder and head of search engine optimization (SEO) at Loopex Digital, suggested a better approach for those using automated tools. “If you want to save time writing emails with automated tools, the better approach is feeding ChatGPT or whatever platform you’re using your own emails and writing samples,” she said.
Training systems on personal writing samples prevents them from falling back on generic phrasing, according to Harutyunyan. She advised never sending generated emails without reading and editing them first, noting that tools work best when they learn individual voices.
Supportive language represents another indicator of machine generation. ChatGPT and other language models use supportive phrases that appear scripted. Expressions like “kindly let me know” or “don’t hesitate to reach out” show up constantly in machine generated text. Experts warn that large language models (LLMs) use these phrases 287 percent more often than actual people do.
George El-Hage, Founder and CEO of digital business card provider Wave Connect, commented on the quality problem. “Email offers have increased 55% in the last year or two, but the quality hasn’t followed. We call it ‘soft spam,'” he said.
He explained that messages look legitimate initially but are generated from broad prompts and miss key details. Wave Connect now assigns every incoming message a relevancy score based on predefined email parameters. This approach cut their email reading time by half and helped focus on genuine offers instead of AI generated content.
The proliferation of artificially generated emails reflects broader adoption of AI writing tools across business communications. While these technologies promise efficiency gains, they create new challenges for recipients trying to identify substantive correspondence among automated outreach.
Detection systems represent one response to the rising volume of machine generated messages. However, as AI writing tools become more sophisticated, the arms race between generation and detection technologies is likely to continue evolving. Companies may need increasingly advanced filters to separate genuine communications from automated bulk messaging.
The trend also raises questions about email effectiveness as a business communication channel. If most recipients develop systems to filter automated messages, the value of AI generated outreach may diminish over time, potentially forcing businesses to reconsider their communication strategies.
