Blogging with Ais is an area fraught with danger. AI text on a page can get you seriously demoted in the world of page ranking and website domain ratings. There are very simple concepts plugged in to the way that AIs respond to questions and statements and the similarities exist in both Gemini and ChatGPT.
These tell-tale signs include simple, analyzable habits such as:
- Bullet points (yes this is sort of a joke) and their abundance. AIs repeatedly just churn out answers to questions with a series of bullet-points and then a very hazy conclusions with a title marked 'conclusion' that can make their responses very easy to spot. This is just as much an error on behalf of the blogger as it is the AI as the blogger can ask for responses or pose certain frames of questioning to avoid these tell-tale signs.
- Response length. Ais are not just AIs, they are bridled with all sorts of rules and regulations and artificial randomness and central to this is the length of response. ChatGPT almost always gives exactly the same length of response and Gemini, whilst it varies a little more, is not that much different.They do not want AIs writing dissertations or CRM systems in a single response because of the energy that might be required and the legal or other problems these situations may cause.
- Rounding-up at the end of every response. It seems that ChatGPt has to give an epilogue of every response rendered which makes the reading of a page created from ChatGPT responses very tiring and uninteresting to read.
Conclusion:
No, just kidding.
There are several other telltale signs that can give away ChatGPT-written copy to Google's algorithms:
-
Repetitive Patterns: ChatGPT tends to repeat phrases or sentence structures, especially when generating longer texts. Google's algorithms are trained to detect such repetitive patterns, which can be a sign of low-quality or AI-generated content.
-
Generic Language: ChatGPT often uses generic and formulaic language, lacking the nuance and originality of human writing. This can make the content feel bland and lacking engagement for readers, which Google might interpret as a sign of lower quality.
-
Lack of Depth: ChatGPT tends to provide surface-level information without delving into the complexities or nuances of a topic. This lack of depth can make the content less informative and valuable to users, which could negatively impact its ranking in Google search results.
-
Inconsistent Style: ChatGPT might struggle to maintain a consistent writing style throughout a piece of content. This can lead to abrupt changes in tone, vocabulary, or sentence structure, which can be jarring for readers and signal to Google that the content might not be well-crafted.
-
Unnatural Phrasing: Despite its impressive language generation capabilities, ChatGPT can sometimes produce sentences that sound awkward or unnatural to human readers. This can be due to limitations in its understanding of context, nuance, and idiomatic expressions.
-
Watermarking: OpenAI is actively researching watermarking techniques to embed subtle patterns or signals in ChatGPT-generated text. These watermarks would be invisible to humans but detectable by algorithms, allowing Google to identify AI-generated content.
-
AI Detection Tools: There are emerging AI detection tools that can analyze text and assess the likelihood of it being generated by AI models like ChatGPT. Google could potentially integrate such tools into its algorithms to identify and filter out AI-generated content.
It's important to note that these are just potential indicators, and Google's algorithms are constantly evolving. There's no foolproof way to determine whether a piece of text was written by ChatGPT, but by being aware of these telltale signs, you can make informed decisions about the content you create and consume.