It might seem hard to imagine a belief-speaking consultancy process rendering useful outcomes, even with the mechanisms as described in an earlier blog post in place: a method to enable the expression of opinions without fear of judgment (“Interdemocracy”), a structured approach to extraction that ensures meaningful outcomes (“Wisdom of Crowds”), and a system for aggregating and interpreting the resulting input (AI). So my wife and I organized a simulation. We asked some of our students to formulate their answers to a relevant question. We then asked ChatGPT to aggregate and interpret the answers. Below you’ll find the outcome of the simulation.
The question
Freedom of speech is a hot button topic on the moment. President Trump signs executive orders to restore his interpretation of freedom of speech (for instance his executive order dated January 20, 2025). European Commissioner Henna Virkkunen gave her interpretation during her confirmation hearing: “freedom of speech is one of our fundamental rights in Europe and we want to make sure that we are protecting this important right for all our citizens. But, of course, then there has to be also balance”.
My wife and I chose to present the issue to our students according to the mechanisms, within a frame rooted in their daily lives. We began with a brief introduction, stating that while freedom of speech is not absolute, the boundaries of this freedom are not always clear. Then we asked them for their take on it: “What is the limit of what can be said in the classroom? Why do you think so?”
The answers
• “I believe that the limit depends on the subject. One must always be aware that everything we say may be used somewhere someday. We must always remember what we say and to whom, and not trust many people—only those closest to us.”
• “If the class is close-knit and does not judge one another, then one can talk about practically anything, as long as it does not hurt others. I think this way because our class is close-knit, and we can talk freely with each other.”
• “The limits of what can be said in class depend on the context of the conversation and the rules in place in the class. I believe we should definitely avoid hate speech, offensive language, and be mindful that our words can hurt someone.”
• “I think crossing the line means hurting someone's feelings with a comment that lacks meaningful criticism. If I want to convey unpleasant but true criticism to someone, I believe I am not crossing the line.’
• “I think one must be aware of what one says and to whom because everyone may interpret what we say differently—one person may feel offended, while another may not pay attention to it. Everyone has the right to express their opinion as long as they do not cross a certain limit.”
• “The limit of speaking in class depends on what we are talking about. In my opinion, anything can be said as long as it does not negatively affect someone's feelings.’
• “The limit of what can be said in class is quite broad because when speaking in front of the class, especially in the presence of a teacher, one must be careful with their words to avoid offending anyone or causing an argument. If the class and the teacher are close-knit, I think practically any topic can be discussed.”
• “I think the most important limits of speaking in class involve thinking first and then speaking while maintaining respect. One must consider the sensitivity of others and whether the words spoken are truthful. When speaking, not just in class, we should remain courteous and take into account what others are saying without interrupting, shouting over them, or behaving in a way that causes discomfort to another person.”
• “The invisible limit of what can be said in class is very difficult to define due to its flexibility and the fact that it is not visible, as there are no strict rules on what can and cannot be said. In my opinion, we can say anything that does not harm another person; we can have different opinions and discuss them, but the limit begins where we cause harm or hurt someone. The problem is that this is often not visible either.”
ChatGPT’s interpretation
We fed the students' answers to ChatGPT. We asked it to provide pattern detection, clustering, outlier identification, aggregation of insights, sentiment analysis, and quality control. It provided us with the following output:
“Pattern Detection
Recurring themes across the responses:
• Context dependency: Many responses emphasize that what can be said depends on the topic and situation.
• Respect for others: Avoiding offensive language, hate speech, and hurting others is a common concern.
• Freedom of speech with limits: Several responses acknowledge the right to express opinions but stress the need for limits.
• Class dynamics: The relationship between classmates and the teacher determines the openness of discussion.
• Awareness of consequences: A common theme is the idea that words can be misinterpreted or used against someone later.
Clustering
1. Context & Awareness (Responses 1, 3, 5, 6, 9)
• Emphasis on adjusting speech based on context, being mindful of what is said, and ensuring that words do not cause harm.
2. Respect & Sensitivity (Responses 2, 4, 7, 8)
• Focus on mutual respect, not offending others, and considering classmates’ feelings.
3. Class Dynamics & Limits (Responses 2, 7, 9)
• The ability to discuss topics freely depends on how close-knit the class is.
4. Freedom vs. Limits (Responses 5, 6, 9)
• Acknowledging that while people should be able to express themselves, limits exist where speech becomes harmful.
Outlier Identification
• Response 4: This is unique because it introduces a distinction between senseless hurtful comments and justified but unpleasant criticism.
• Response 9: Stands out due to its focus on the "invisible" nature of limits and the difficulty of defining them.
Aggregation of Insights
• Speech in class should be context-aware, respectful, and sensitive to others’ feelings.
• Class atmosphere matters: A close-knit class can discuss more topics openly.
• Limits exist but are flexible: They depend on the situation, the people involved, and the nature of the speech.
• Words have consequences: Students recognize that what they say can be misinterpreted or used against them.
Sentiment Analysis
• Neutral-to-positive: Most responses acknowledge the importance of free expression but advocate for responsible speech.
• Concerned/Cautious: Some responses (1, 3, 5, 9) express a sense of caution about how words may be misused.
• Optimistic: Responses (2, 7) highlight a positive classroom environment where open discussion is possible.
• Critical/Evaluative: Response 4 is unique in distinguishing between harmful speech and constructive criticism.
Quality Control
• Clarity & Coherence: Some responses (e.g., 7, 9) have long sentences that could be more structured for clarity.
• Repetitiveness: Some responses reiterate similar ideas in different words.
• Depth of Thought: Response 4 stands out for its nuanced take on criticism versus offensive speech.”
Conclusion
The simulation outcomes look promising. What needs to be established next is the form in which the AI output should be presented to the Youth Resilience Council (YRC) so that the YRC can start to formulate its recommendations.
We cannot wait to test the belief-speaking consultative process in a large-scale pilot that also includes an actual YRC.
