Overview :
Here are 10 very useful tips to help you to spot the use of AI when reviewing documents. This new critical communication skill is becoming essential.
“ChatGPT is one of those rare moments in technology where you see a glimmer of how everything is going to be different going forward.” ~ Aaron Levie
Communication skills to address AI:
At the end of last year, I received a very personal thank you message from another trainer generated by ChatGPT. Communication alarm bells rang out in my head. The future was here!
AI is now evolving at lightning speed. These advanced AI systems are quickly becoming increasingly clever at mimicking our human writing styles, making it very challenging for readers, managers and clients to discern whether an article has been crafted by a human or a machine.
Luckily, there are some subtle cues that can help you to identify if a document has been written by platforms like ChatGPT. In this article, I explore ten very useful tips to help you to spot the fingerprints of artificial intelligence in written content:
- Overly impressive writing:
Only 9,6% of South Africans are English mother tongue speakers. Professional written English proficiency levels are sadly generally lacking in the wider population, even in the higher echelons of the corporate world. So, if a document is written with a higher standard of written English than the writer seems capable of, this may well be a very easy telltale sign that the writer has enlisted the help of AI. It is worth studying the writing styles of your stakeholders and notice if they suddenly change or become elevated.
- Lack of Personal Touch:
One of the most obvious indicators that an article may be the product of AI is the absence of a genuine personal touch. Humans often inject their subjective experiences, emotions and unique perspectives into their writing. If an article feels sterile / flat and lacks a distinct human emotional touch, it may be the creation of AI.
- Unusual Sentence Structures:
AI, while incredibly clever, may sometimes produce sentences with odd structures or clumsy word choices. Keep an eye out for sentences that seem overly complex, or lacks the natural relaxed flow associated with human communication.
- Inconsistencies in Tone:
Humans generally maintain a consistent tone throughout their writing, reflecting our unique particular styles and voices. In contrast, AI models like ChatGPT might exhibit fluctuations in tone or writing style within the same piece. Pay attention to sudden shifts in language, formality, or tone. These inconsistencies may reveal the artificial nature of the content.
- Lack of Current Events Awareness:
AI models lack real present-time awareness and are typically trained on data up to a certain point. If a document fails to address recent events, developments, or changes in a given field, it might signal that the content is generated by AI. ChatGPT’s last update was in January 2022. Its databanks only include data up to that point.
“It takes intelligence to recognise intelligence, or the lack of it.”
~ Arne Klingenberg
- Generic Nature of Examples:
While ChatGPT is trained on an extensive databank, it may struggle to produce nuanced and specific examples. If an article relies heavily on generic or stereotypical examples without delving into unique scenarios or insights, it could be indicative of AI authorship. Humans draw from layered personal experiences, cultural references, and specific insights to enrich their content with authenticity.
- Lack of Bias or Subjectivity:
Human writers generally infuse their work with personal biases, solid opinions and subjective viewpoints. AI models like ChatGPT are specifically designed to be neutral and objective, which can result in content that lacks the inherent subjectivity found in human writing. If an article appears overly impartial or devoid of personal opinion, it may be a strong clue that artificial intelligence played a role in its creation.
- Unnatural Conversational Flow:
Despite recent improvements in conversational AI, there are still some instances where the flow of information in a document feels forced or unnatural. If the document’s conversational flow feels rigid or lacks the ebb and flow typical of human interaction, it could be a sign of AI crafting.
- Contextual Understanding:
Pay attention to the article’s overall ‘bigger picture’ contextual understanding. While AI can generate coherent sentences, it may struggle with maintaining a deeper understanding of the broader context or themes. If an article appears to lack a holistic grasp of the subject matter or fails to delve into more nuanced aspects, it might be an indication of AI-generated content.
- AI detector software:
Identifying machine-generated content is currently possible, but it is not foolproof. The appropriate software tool for you depends on your specific goals and the general context in which you’re working. The following article from Forbes unpacks the current AI detector software you have available:
Link: https://www.forbes.com/sites/technology/article/best-ai-content-detector-tools/?sh=5ee3f4a15d5e
Conclusion:
As AI language models continue to become more mainstream and abused by humans, one of the new critical skills we need to develop is to sharpen our assessment skills to distinguish between human and machine-generated content. By being able to recognise the nuances of language and writing, you can navigate the evolving landscape of AI-generated content with greater discernment.
Happy reading!
