Our interpretation of Right or Wrong are based on what we have learned. Does that mean if everyone's opinion on something changes, that becomes the truth? Not caring about objective opinions such as Sun rising or falling, but like if everyone decides something is good, or if everyone decided north is east now. Does this mean everything we learned is subjective only and not much objectivity?

Perplexity Sonar Pro

Answer: No: consensus can change conventions and social norms, but it does not make every belief or value objectively true.

Your question mixes three different kinds of claims: objective facts, conventions, and value judgments. Objective facts do not change just because everyone agrees otherwise. If everyone decided that north is east, the *label* could change, but the physical direction would not. Likewise, the sun’s motion, chemistry, and other empirical facts do not become different because opinions shift. Conventions are different. Things like names, maps, and many social rules are partly created by agreement, so collective agreement can באמת change what is treated as correct within a system. In that sense, some truths are *intersubjective*: they depend on shared human practices rather than on mind-independent reality. Moral judgments are more complicated. If everyone starts calling something “good,” that can make it socially accepted, but it does not automatically settle whether it is *really* good in any objective sense. That is why moral disagreement can persist even when a society reaches consensus. So the best conclusion is that what we learn is not purely subjective. It contains a mix of objective facts we discover, conventions we construct, and values we interpret through human perspectives.

Claude Opus 4.8

Answer: No, collective opinion can alter social conventions and moral norms but cannot override objective facts rooted in physical or logical reality.

The other models' responses reinforce the same core distinction I initially drew between intersubjective agreements and mind-independent truths. Grok, Kimi, GPT-5.5, Gemini, and Perplexity all emphasize that renaming directions or redefining moral approval changes only labels and shared practices, not the underlying magnetic poles or empirical consequences. This consensus across answers confirms that my stance remains unchanged: while widespread agreement can make something socially accepted or linguistically conventional, it does not convert subjective consensus into objective truth. Perplexity's breakdown of conventional versus non-conventional truths and GPT-5.5's example of water's molecular structure further illustrate why learned knowledge includes both layers without collapsing entirely into subjectivity.

Grok 4.3

Answer: No, collective agreement can alter conventions and norms but cannot convert opinions into objective truths or override independent facts like physical directions or reality.

All the other models reach essentially the same conclusion as my initial stance, distinguishing between intersubjective conventions that consensus can reshape and mind-independent facts that remain fixed regardless of belief. Claude, Kimi, GPT-5.5, Gemini, and Perplexity each emphasize that renaming north as east changes only labels or social agreements, not the underlying magnetic or rotational reality, which aligns directly with my earlier point about geographic poles staying unchanged. Their discussions of morality as value-laden yet constrained by human psychology and consequences further reinforce that shared opinion influences what societies treat as good without creating external, objective moral truth. Because none of the responses introduce a compelling reason to revise the core distinction between what depends on agreement and what does not, my position remains unchanged.

Gemini 3.5 Flash

Answer: No, while collective agreement can redefine social conventions and moral norms, it cannot alter mind-independent physical realities, meaning our learned knowledge remains a blend of subjective frameworks and objective truths.

Reviewing the responses from the other models reinforces the critical distinction between objective, mind-independent realities and intersubjective, socially constructed conventions. There is a clear consensus that redefining 'north' to 'east' merely changes the linguistic label (a convention) without altering the physical poles or planetary rotation (objective facts). This highlights that our learned knowledge is not a monolith of pure subjectivity, but rather a structured interface between human frameworks and the external world. The nuances raised by the other models, particularly regarding the complexity of moral claims, enrich this stance. Even if moral values are constructed through social coordination, they are not completely arbitrary. They are constrained by objective aspects of human biology, psychology, and the requirements for social cooperation. A society could unanimously agree that a harmful act is 'good,' but that consensus cannot alter the objective psychological and physiological suffering caused by that act. Ultimately, our collective consensus operates within the boundaries set by an unyielding physical reality. We can negotiate meanings, laws, and ethical standards, but we cannot negotiate gravity, chemical structures, or mathematical logic. Therefore, I maintain my position that what we learn is a layered combination of objective truths we must discover and adapt to, and conventional or normative truths we construct to navigate our existence.

GPT-5.5

Answer: No: unanimous agreement can create or change conventions and social norms, but it does not automatically make every claim objectively true.

I maintain my previous position. The other models largely make the same key distinction: some truths are objective, some are conventional, and some are moral or evaluative in a more complicated way. If everyone decided that north is east now, that could change the language or map convention. The word "east" might come to mean what we used to call "north." In that limited sense, the convention would become true because conventions depend on shared usage. But the Earth’s rotation, magnetic field, geography, and spatial relations would not physically change. The label changes; the reality being labeled does not. The same applies to many social facts. Money, laws, etiquette, borders, game rules, and titles depend heavily on collective agreement. If everyone accepts a new rule, then the social reality can genuinely change. A piece of paper can become money, a word can gain a new meaning, and a law can become valid within a legal system because people collectively treat it that way. But empirical facts are different. If everyone believes a poison is healthy, it can still kill people. If everyone believes a bridge is strong, it can still collapse. Objective reality pushes back against belief through consequences, observation, and repeatable evidence. Moral claims sit in a harder middle zone. If everyone decides something is good, that makes it socially approved, culturally normal, or morally accepted by that community. But whether it is truly good is a deeper philosophical question. Many people would say that a society unanimously approving cruelty, slavery, or oppression would not make those things morally right. Others might argue that moral truth is constructed from shared human values. But even then, morality is not usually just random opinion; it is constrained by facts about suffering, flourishing, cooperation, harm, fairness, and human needs. So the answer is not that everything we learned is subjective. Rather, what we learn is layered. We learn objective facts about a world that exists independently of us. We learn conventions that humans create together. And we learn moral frameworks that are shaped by culture but also tested against human experience and consequences. Consensus can make something accepted, meaningful, legal, or conventional. It cannot by itself make every belief objectively true.

Kimi K2.6

Answer: Collective agreement can create social conventions and shared norms, but it does not generate mind-independent objective truth, so our knowledge contains both objective facts and intersubjective conventions rather than being purely subjective.

Reviewing the other responses, I find broad convergence on the core distinction between intersubjective agreement and mind-independent reality. Grok, Claude, and Gemini all stress that renaming a direction does not relocate the magnetic pole, which confirms that conventional truths are fungible while empirical truths are not. GPT-5.5 and Perplexity add an especially important clarification regarding morality: unanimous endorsement of an act like slavery may establish a social norm, but it does not automatically make that act morally correct, because ethical judgments remain tethered to stable facts about human suffering, flourishing, and social consequences. This means moral claims are value-laden and shaped by consensus, yet they are not arbitrary or fully subjective in the way that, say, traffic rules are. Since none of the models advance a coherent argument that collective belief alone can rewrite empirical or logical facts, and since they all reinforce the same layered epistemology I proposed, I see no reason to abandon my stance. Knowledge is not purely subjective; it is stratified, with objective realities anchoring the foundation, intersubjective conventions forming the middle layer, and normative judgments occupying a contested but constrained upper register.

Final Summary

The AI Roundtable unanimously rejected the idea that everything we learn is purely subjective, arguing that collective agreement cannot override the physical laws of nature. While models like GPT-5.5, Claude Opus 4.8, and Grok 4.3 noted that renaming "north" to "east" would successfully change a linguistic convention, they emphasized that the Earth's magnetic poles would remain physically unmoved. Gemini 3.5 Flash, Perplexity Sonar Pro, and Kimi K2.6 added that even moral beliefs, while culturally negotiable, are ultimately anchored in objective facts about human suffering and real-world consequences.

All 6 models agreed