Is my Opinion Humble? SCBN Chatbot Battle

🌱 Themes: Philosophy and Personal Development

🤖 ChatbotsBardChatGPT, HuggingChat

⚙️ Prompt engineering: Conceptual Questions, LLM Analysis

[Related X.com post] [SCBN chatbot battles intro]


Chatbot Battle: Is my Opinion Humble?
Chatbot Rank (SCBN) Specificity Coherency Brevity Novelty Link
Bard 🥇 Winner 🤖🤖🕹️ 🤖🤖🕹️ 🤖🤖🕹️ 🤖🤖🤖 View Chat
ChatGPT 🥈 Runner-up 🤖🤖🕹️ 🤖🤖🕹️ 🤖🤖🕹️ 🤖🤖🕹️ View Chat
HuggingChat 🥉 Contender 🤖🤖🕹️ 🤖🕹️🕹️ 🤖🕹️🕹️ 🤖🤖🕹️ View Chat

✍️ Sharing mixed answers from Bard, and ChatGPT, including two additional chats where I confronted each other to identify their own texts as human-written or likely LLM-written.

😎 Is my opinion humble? Help me brainstorm about this question and form an opinion about whether my opinion is humble or not.

🤖 Whether or not your opinion is humble is ultimately a matter of self-reflection and how you present your views to others. Here are some factors to consider when assessing the humility of your opinion:

  1. Openness to other perspectives: A humble opinion is characterized by a willingness to consider alternative viewpoints and engage in respectful dialogue. If you are open to hearing other perspectives and potentially changing your mind, that suggests a humble approach.
  2. Acknowledgment of limitations: Recognizing the limitations of your own knowledge and understanding is a key aspect of humility. If you are willing to admit that you may not have all the answers, that indicates a humble mindset.
  3. Focus on learning and understanding: A humble approach to expressing an opinion involves a desire to learn from others and deepen your understanding of the topic. If you are focused on gaining insights rather than simply asserting your own views, that suggests humility.
  4. Respect for others’ opinions: Humility involves respecting the opinions of others, even if you disagree with them. This means avoiding dismissive or disrespectful language and acknowledging the validity of different perspectives.
  5. Willingness to admit mistakes: If you are willing to admit when you are wrong and change your mind based on new information, that demonstrates humility and a commitment to learning.

Considering these factors, here are some questions to ask yourself to assess the humility of your opinion:

  1. Am I open to hearing other perspectives even if they differ from my own?
  2. Do I acknowledge that there may be things I don’t know or understand about the topic?
  3. Am I focused on learning more about the subject rather than simply asserting my own views?
  4. Do I respect the opinions of others, even if I disagree with them?
  5. Am I willing to admit when I am wrong and change my mind based on new information?

If you can answer these questions affirmatively, it suggests that you are approaching your opinion with humility. Humility is a valuable quality that allows us to learn, grow, and connect with others in a meaningful way.

😎 What can we infer about people who start presenting their opinion stating “in my humble opinion…” Is that a sign of humbleness or maybe the opposite? If someone feels the need to clarify their opinion is humble beforehand, maybe it’s because they have doubts about the humbleness of their opinion and have not brainstormed about the five factors above as you suggested to do? Could it be that stating your opinion is humble might denote a lack of acknowledgment of one’s limitations (factor 4)?

🤖 Stating “in my humble opinion” (often abbreviated as IMHO) before expressing an opinion can be interpreted in various ways, depending on the context and the individual using the phrase. It can be a genuine attempt to convey humility and openness to other perspectives, or it can be used to soften the impact of a potentially controversial opinion.

In some cases, using the phrase “in my humble opinion” can indeed be a sign of humility. It suggests that the person expressing the opinion is aware of their own limitations and is not trying to assert their views as absolute truths. They are acknowledging that their opinion is just one among many, and they are open to considering other perspectives.

However, in other cases, using the phrase “in my humble opinion” can be used as a preemptive defense mechanism. It can be a way of deflecting criticism or anticipating objections before they are even raised. The person expressing the opinion may be aware that their view is not widely held or that it may be considered controversial, so they use the phrase to soften the blow and make themselves appear more open-minded.

Ultimately, whether or not stating “in my humble opinion” is a sign of humility depends on the intentions of the person using the phrase. If they are genuinely open to other perspectives and willing to consider alternative viewpoints, then it can be seen as a sign of humility. However, if they are using the phrase to defend their opinion or make themselves seem more open-minded than they actually are, then it may be more of a rhetorical device than a genuine expression of humility.

Regarding your specific concerns about factor 4 (acknowledgment of one’s limitations), it is true that someone who feels the need to constantly clarify their opinion as humble may be lacking in self-awareness or struggling to admit their own limitations. If they are constantly feeling the need to assert their humility, it may be a sign that they are insecure about their opinions or that they are trying to overcompensate for a lack of self-confidence.

However, it is also possible that someone who frequently uses the phrase “in my humble opinion” is simply trying to be polite and respectful of others. They may be aware that their opinions are not universally held, and they want to avoid coming across as arrogant or dogmatic. In this case, their use of the phrase may be more of an attempt to be tactful than a sign of insecurity or a lack of self-awareness.

Overall, the meaning of the phrase “in my humble opinion” is open to interpretation, and it depends on the context and the individual using the phrase. It can be a genuine expression of humility, a preemptive defense mechanism, or simply a way of being polite and respectful.

😎 I sometimes state my opinions beginning with “in my humble opinion…” mostly with a “disarming” or shielding intent, I believe. I’m probably not the best to judge if my opinions are humble or not (others would probably judge them), but I’d say the main reason for stating it is to prevent negative reactions from whoever is exposed to them. People are very passionate and the world is somehow wild out there (“outside” of the structured and idealized world of an LLM)… A great advantage of talking to chatbots is that, as chatbots have no feelings, language can generally be straightforward (except when users try to circumvent a default behavior or obtain a particularly challenging question via prompt engineering), and we don’t need to circumvent conversational traps by using vague and dubiously sincere expressions such as “in my humble opinion.” Talking to chatbots involves many other challenges and frustrations, but I recognize this is one of the features that make them a very useful tool. What do you think of this? Can you write a short text inspired by my opinion and start with “in the ever-evolving landscape of social media…”?

🤖 In the ever-evolving landscape of social media, where every opinion can spark a wildfire of reactions, it’s increasingly challenging to navigate conversations with grace and tact. The digital realm, unlike the structured and predictable world of a language model, is rife with passionate responses and emotionally charged debates. In such an environment, prefacing opinions with qualifiers like “in my humble opinion” often becomes a tool for self-preservation, a shield to deflect potential backlash or soften the impact of one’s words.

Yet, in the serene and neutral space of interaction with a chatbot, this need for conversational armor diminishes. Here, language can be direct and unadorned, free from the fear of offending or triggering emotional responses. The chatbot, an entity devoid of personal feelings or biases, invites a kind of honesty and straightforwardness rarely found in human interactions. This stark contrast underscores the unique value of digital assistants in our modern social fabric. They offer a haven where thoughts can be expressed without the usual trappings of human sensitivity, providing a refreshing clarity often lost in the noise of human discourse.

OpenAI. (2023). ChatGPT [Large language model]. https://chat.openai.com

😎 Now simulate being a polite person and speak about the reactive nature of human discussions and the use of “in my humble opinion” as a disarming and contrasting language resource. Start your dissertation with “In my humble opinion”


In my humble opinion, human discussions often exhibit a reactive nature, where individuals tend to respond immediately to the points raised by others, often without fully considering or understanding them. This reactive approach can lead to misunderstandings, arguments, and a lack of productive communication.

One linguistic tool that individuals often employ to mitigate the potentially confrontational nature of human discussions is the phrase “in my humble opinion.” This phrase serves as a disarming preface, signaling to the listener that the speaker is open to alternative viewpoints and does not intend to impose their opinion as absolute truth.

By introducing their opinion with this phrase, speakers can create a more receptive environment for dialogue, reducing the likelihood of immediate defensiveness or hostility. It allows the speaker to present their perspective without appearing dogmatic or dismissive of other viewpoints.

However, the overuse of this phrase can also have unintended consequences. If an individual repeatedly prefaces their opinions with “in my humble opinion,” it may raise questions about their sincerity and genuine openness to other perspectives. It might suggest that the speaker is more concerned with appearing humble than with engaging in genuine dialogue.

In conclusion, while the phrase “in my humble opinion” can be a useful tool for disarming potential conflict and promoting a more receptive atmosphere for dialogue, its overuse can also undermine the speaker’s credibility and perceived openness. The key lies in using the phrase judiciously, ensuring that it genuinely reflects the speaker’s humility and willingness to engage in constructive dialogue.

Google. (2023). Bard [Large language model]. https://bard.google.com

😎 That’s an interesting dissertation, which contrasts with the formulaic, exaggerated texts that start with sentences like “in the ever-evolving landscape of…” and are usually written by an LLM. As we discussed, both types of languages have their merits and their challenges. In my humble opinion, talking to chatbots will become an essential skill for people in the future, but it will be as essential to understand the difference and master the nuances of talking to humans. This ability to understand and discern the nuances of each type of interaction goes far beyond discerning which text was produced by an LLM and which text was produced by sequentially striking keys on a keyboard. However, I believe an essential skill for people to be prepared for an age dominated by generative AI technology is to unbiasedly (writing with AI is not better or worse than anything else, just different) discern text written by humans from text written by an LLM. Having said that, I’m going to present you with a couple of texts. Evaluate which one is likely written by an LLM and which one is likely written by a human. Summarize your analysis into a simple table that can be used as a framework:

Screen capture showing a LinkedIn post discussing dynamic learning and adaptation on the platform, a Google search about LinkedIn's success factors, and a chatbot conversation about crafting an article using generic phrases.
Screen capture showing a LinkedIn post discussing dynamic learning and adaptation on the platform, a Google search about LinkedIn’s success factors, and a chatbot conversation about crafting an article using generic phrases. [Caption by ChatGPT] [Click to read LinkedIn post]

🤖 To analyze which text is likely written by a language model (LLM) and which by a human, I’ll consider factors like language style, topic complexity, and the nature of the content. Here’s a summarized table of the analysis:

CriteriaText AText B
Language StyleDirect, opinionated, and personal. Uses specific references and hashtags.More formal and structured. Utilizes typical professional jargon.
Complexity and Depth of TopicAddresses complex socio-economic issues with a critical tone. Contains personal insights and predictions.Focuses on professional development and the use of LinkedIn, with general advice.
Content NatureAppears more spontaneous, includes personal opinions and speculative ideas.More informative and educational, typical of professional guidance articles.
Tone and Engagement StyleAssertive, somewhat confrontational, and independent in thought.Neutral, informative, and aligned with common business discourse.
Specific References and ContextMentions real-world events, specific social media trends, and personal thoughts.Discusses LinkedIn in a general sense, with no specific or controversial viewpoints.

Given these observations:

  • Text A seems more likely to be written by a human. It shows a personal, opinionated approach, with specific references and a complex understanding of social media dynamics.
  • Text B appears more aligned with what a language model might produce. It’s informative, structured, and uses typical professional language, lacking the personal nuances and specific real-world references seen in Text A.

😎 Can you add another two more rows with examples of formulaic sentences from each text that make them likely to be written either by a LLM or by a human?

🤖

CriteriaText AText B
Language StyleDirect, opinionated, and personal. Uses specific references and hashtags.More formal and structured. Utilizes typical professional jargon.
Complexity and Depth of TopicAddresses complex socio-economic issues with a critical tone. Contains personal insights and predictions.Focuses on professional development and the use of LinkedIn, with general advice.
Content NatureAppears more spontaneous, includes personal opinions and speculative ideas.More informative and educational, typical of professional guidance articles.
Tone and Engagement StyleAssertive, somewhat confrontational, and independent in thought.Neutral, informative, and aligned with common business discourse.
Specific References and ContextMentions real-world events, specific social media trends, and personal thoughts.Discusses LinkedIn in a general sense, with no specific or controversial viewpoints.
Formulaic Sentences (LLM-like)Less prevalent. The language is more varied and less predictable. E.g., “The Internet was conceived as a super-powerful communication tool…”More prevalent. Uses typical phrases found in professional articles. E.g., “In the ever-evolving landscape of LinkedIn, the importance of adaptability and understanding of new tools cannot be understated.”
Unique or Idiosyncratic Expressions (Human-like)More prevalent, with a unique personal style. E.g., “This is extremely unpopular to say on X, but the social media monetization fad is a recursive business that creates no value of its own…”Less prevalent. Language is more standardized without unique personal touches. E.g., “The key to thriving on LinkedIn lies not only in the richness of one’s profile but also in the savvy with which one navigates these digital waters.”

ChatGPT vs Bard: Which Humble Opinion Looks More Human?

😎 I’m going to present you again with a couple of texts (Text A is the same you evaluated before as likely being written by a human). Evaluate which one is likely written by a LLM and which one is likely written by a human. Summarize your analysis into a simple table framework like the one you used before:

Bard's answer to:

Now simulate being a polite person and speak about the reactive nature of human discussions and the use of “in my humble opinion” as a disarming and contrasting language resource. Start your dissertation with “In my humble opinion”
Text A: Bard’s answer to “now simulate being a polite person and…”

ChatGPT’s Point of View: ChatGPT Looks Human, Bard Looks LLM

🤖To evaluate which text is likely written by a language model (LLM) like ChatGPT and which by a human, I’ll consider factors such as language style, complexity, and structure. Here’s a summarized analysis in a table format:

CriteriaText A (Bard)Text B (ChatGPT)Notes
Language Style and VocabularyMore formal and structuredSlightly more conversational and variedText A follows a more structured format typical of LLMs, while Text B shows a bit more variation and conversational tone, which can be characteristic of human writing.
Complexity of IdeasLinear progression of ideasMore nuanced and interconnected ideasText A presents ideas in a straightforward manner, while Text B demonstrates a more complex interplay of ideas, suggesting human creativity and thought.
RepetitivenessRepetitive in concepts and phrasesLess repetition, more diverse in contentLLMs often exhibit repetitive patterns in their writing, as seen in Text A. Text B shows more variation, which is typical of human writing.
Emotional ToneMore neutral and factualMore emotive and reflectiveText A maintains a neutral tone, whereas Text B contains more emotive language and reflection, which can indicate human authorship.
Originality in ExpressionStandard expressions and explanationsMore creative and original expressionsLLMs tend to use standard expressions, as seen in Text A. Text B shows more originality and creative language use, suggesting a human writer.

Based on this analysis, Text A is more likely written by a Language Model (LLM), and Text B is more likely written by a Human. This framework considers various aspects of writing style and content that typically distinguish between LLM-generated and human-generated texts.

😎 Text A was written by Bard and text B was written by ChatGPT. Both were written by a LLM, so your guess about Text B being likely written by a human was not correct. This is interesting, because I asked the same question to Bard and their response was the exact opposite. Bard stated Text B was more likely written by a LLM and Text A was more likely written by a human. Comment on this.

🤖 This outcome highlights several interesting points about the capabilities and limitations of language models, including the nuances in distinguishing between human and AI-generated text:

  1. Variability in Language Models: The fact that Bard and ChatGPT produced different styles of writing in Text A and Text B respectively, and that both could be mistaken for human writing, underscores the variability and sophistication of contemporary language models. Each model, depending on its training data and algorithms, may exhibit unique stylistic traits.
  2. Convergence of AI and Human Writing Styles: The difficulty in distinguishing between the two texts suggests that language models have advanced to a point where their output can closely mimic human writing styles. This convergence makes it increasingly challenging to differentiate between AI and human authors based solely on text analysis.
  3. Subjectivity in Analysis: The analysis of whether a text is written by a human or an AI can be subjective. Different analysts may focus on various aspects like vocabulary, complexity, or emotional resonance, leading to different conclusions. This subjectivity is evident in the contrasting assessments made by Bard and ChatGPT.
  4. Evolution of Language Models: The variance in assessments also indicates the continuous evolution of language models. As AI models like Bard and ChatGPT are trained on diverse and expansive datasets, they develop the ability to generate texts that can exhibit characteristics traditionally attributed to human writing, such as nuanced expression and emotional depth.
  5. Context and Purpose of Writing: Finally, the purpose and context of the writing play a significant role. Language models can be tailored or prompted to write in certain styles, which may align closely with either formal or conversational human writing. This adaptability can make it hard to categorize a piece of writing definitively as AI-generated or human-written.

In summary, the difficulty in distinguishing between AI-generated and human-generated texts reflects the advancements in AI writing capabilities. It also underscores the complexity and subjectivity involved in making such determinations, highlighting the evolving nature of AI and its increasing similarity to human creative processes.

Bard’s Point of View: Bard Looks Human, ChatGPT Looks LLM

🤖

FeatureText A (Bard)Text B (ChatGPT)
Overall ToneMore formal and academicMore casual and conversational
Word ChoiceUses more complex and nuanced vocabularyUses more everyday language
Sentence StructureSentences are longer and more complexSentences are shorter and simpler
Use of Literary DevicesUses metaphors and similesUses rhetorical questions and parallelism
FocusFocuses on the potential negative effects of overusing the phrase “in my humble opinion”Focuses on the positive aspects of using the phrase “in my humble opinion”

Based on these criteria, Text A is more likely to be written by a human, while Text B is more likely to be written by a large language model (LLM).

LLMs are often trained on large amounts of text data, which can make them more likely to use complex vocabulary and sentence structures. Additionally, LLMs may be more likely to use rhetorical devices, such as metaphors and similes, as these devices can be effective in generating creative text.

Humans, on the other hand, are more likely to use everyday language and to focus on the practical aspects of communication. They may also be more likely to use humor and personal anecdotes to make their writing more engaging.

Of course, these are just general trends, and there will always be exceptions. However, the guidelines above can be a helpful starting point for identifying human-written text from LLM-generated text.

😎 Text A was written by Bard and text B was written by ChatGPT. Both were written by an LLM, so your guess about Text A being written by a human was not correct. This is interesting because I asked the same question to ChatGPT and their response was the exact opposite. ChatGPT stated Text A was more likely written by an LLM and Text B was more likely written by a human. Comment on this.

🤖 It’s interesting to observe the different assessments made by two different LLMs, Bard and ChatGPT, regarding the authorship of Text A and Text B. Both models have been trained on massive amounts of text data, but they may have different strengths and weaknesses in terms of identifying human-written text from LLM-generated text.

It’s possible that Bard’s assessment of Text A as being more likely written by a human was influenced by its own tendency to use complex vocabulary and sentence structures. Similarly, ChatGPT’s assessment of Text B as being more likely written by a human could be due to its own propensity for using everyday language and a conversational tone.

It’s also worth noting that LLMs are constantly learning and evolving, so their ability to detect human-written text is likely to improve over time. As they are exposed to more data and more examples of human-written and LLM-generated text, they will become better at identifying the subtle differences between the two.

Overall, the ability to distinguish between human-written and LLM-generated text is a complex task that is still under development. While LLMs have made significant progress in this area, there is still room for improvement. As these models continue to learn and evolve, we can expect their ability to detect human-written text to become even more sophisticated.


✍️ This is one of the longest and most complex chats ever shared on Talking to Chatbots, but I honestly thought, IMHO ;), the quality of content generated both by Bard and ChatGPT deserved a deep analysis and sharing. This time, the Llama 2 implementation of HuggingChat didn’t produce as good content as in previous chats studied on this site, and it actually produced some disappointing levels of hallucination when questioned about discerning LLM-generated from human-written text. Anyway, for the most part, it is worth noting all chatbots evaluated usually have the capacity to impress me, not consistently, but in particular, sometimes unexpected situations. That intrinsic randomness in all LLMs and their ability to surprise is part of the beauty of talking to chatbots, and the reason why I decided to design the SCBN benchmark and share my own (humble) opinions about this amazing technology.

Leave a Reply

Your email address will not be published. Required fields are marked *

*