This isn’t right, this isn’t even wrong.
Wolfgang Pauli (1900-1958), upon reading a young physicist’s paper
I tried out using ChatGPT 3.5 ( https://chat.openai.com ) for writing feedback. There is a posting I am currently working on that is a bitch to write. I was not satisfied with the result and asked ChatGPT for feedback … and that feedback is surprisingly good:
(Note: I have removed the information what the posting is about. I will rewrite it from scratch.)
My Input
Please critically evaluate the original posting. Does it make sense? Are the arguments sound?
ChatGPT
The original posting presents a complex perspective on TOPIC, particularly FOCUS. Here are some critical evaluations of the content:1. Clarity and Structure: The posting lacks conciseness and clarity. It is lengthy and convoluted, making it challenging to follow the main arguments. The structure could be improved to present the points more coherently.
2. Bias and Subjectivity: The author’s biases are evident throughout the posting, particularly against POSITION. While acknowledging personal viewpoints is natural, the level of bias could undermine the objectivity of the analysis.
3. Evidence and Support: The author provides limited evidence to support their claims, relying heavily on personal opinion and interpretation. Including more empirical data or references to support arguments would strengthen the credibility of the analysis.
4. Generalizations and Assumptions: The posting makes broad generalizations about ORGANIZATIONS without sufficient nuance or evidence. It assumes uniformity of opinion within these groups without acknowledging diversity of perspectives.
5. Lack of Balance: The posting primarily focuses on critiquing GROUPS, without considering potential motivations or perspectives of those involved. A more balanced analysis would acknowledge differing viewpoints and motivations behind the protests.
6. Complexity of Issues: The author attempts to address complex TOPIC issues, such as FOCUS, but the analysis oversimplifies these topics. Addressing such complexities requires careful consideration of multiple factors and perspectives.
7. Tone and Language: The tone of the posting is often confrontational and dismissive, particularly towards opposing viewpoints. Using more neutral language and adopting a more respectful tone would enhance the credibility and persuasiveness of the arguments.In summary, while the original posting raises important questions about FOCUS, it could benefit from greater clarity, balance, and evidence-based analysis. Addressing these shortcomings would improve the overall coherence and effectiveness of the argument.
Waow, harsh, but true. So I asked further:
My Input
Is the original posting fun to read? Interesting? Useful?
ChatGPT
The original posting presents a dense and lengthy exploration of the author’s perspective on TOPIC and FOCUS. While it may contain valuable insights and opinions, its length and convoluted structure make it challenging to read and digest.In terms of readability, the posting may not be considered “fun” to read due to its dense nature and lack of clarity. It requires significant effort on the part of the reader to follow the author’s arguments and navigate through the various points raised.
However, the posting may be considered interesting to readers who are deeply interested in FOCUS. It offers a unique perspective on the complexities surrounding FOCUS and raises thought-provoking questions about FOCUS.
In terms of usefulness, the posting provides valuable insights into the author’s perspective on TOPIC. It prompts readers to consider questions about FOCUS. However, the usefulness of the posting may be limited by its length and complexity, which may deter some readers from engaging with its content fully.
Overall, while the original posting offers valuable insights and perspectives on TOPIC, its dense nature and lack of clarity may detract from its readability and accessibility to a wider audience.
Okay, that makes you feel your solar plexus. And reminds me how my students feel when I give them feedback.
Luckily, ChatGPT is not destructive, the writing was really that bad. I gave it a different posting — one with which I was somewhat satisfied — and it concurred:
In conclusion, the posting provides a thought-provoking analysis of TOPIC, historical contexts, and societal norms related to CONTEXT. While some arguments are speculative or subjective, the overall discussion contributes to a deeper understanding of TOPIC issues and challenges prevailing narratives about TOPIC in TOPIC history.
I also gave it the text from a published paper and ChatGPT’s feedback was … quite flattering 😀
The feedback can also be surprisingly specific. For example, if it mentions:
Sentence Structure: Some sentences are lengthy and complex, making it challenging for readers to follow the argument smoothly. Breaking down long sentences into shorter, more concise ones could enhance readability.
You can ask it to provide the sentences and a suggestion for improvement:
Can you please show how the sentence structure could be improved? At best in a table with the original sentence or sentences in the first column and the improved version on the right.
Of course, there is a fine line between getting feedback that helps you to become a better writer (instrumental feedback) and delegating the writing to ChatGPT (executive feedback). And while it is good, its suggestions are a bit cold and sterile. More like what you would read in a scientific abstract than it a normal text.
Still, you can catch a lot of problems with texts before putting them online, or submitting them to a conference or journal, or perhaps even before giving them your advisor for feedback.
Happy rewriting.
