Can text generation be meaningfully called communication without mutual orientation?
This explores whether AI text output counts as real communication when there's no shared attention or mutual uptake between the writer and reader — and what the corpus says is structurally missing.
This explores whether AI text output counts as real communication when there's no shared attention or mutual uptake — and the corpus answers, fairly directly, that something load-bearing is missing. The cleanest framing comes from the argument that communication is social action between people, not information distribution: a genuine exchange does work in a relationship, carries speaker responsibility, and depends on mutual uptake — and AI generates content without any of that relational structure, while the chat interface hides the difference Does AI really communicate or just distribute information?. So the short answer to the question is no — but the more interesting answer is *why*, and the corpus decomposes the 'mutual orientation' you're asking about into several distinct missing pieces.
One piece is the event itself. AI doesn't produce utterances so much as 'event-residue' — text that wears the markers of communication inherited from training data but lacks the underlying event that makes an utterance an utterance. The reader then supplies the missing orientation through interpretive labor, so the exchange has structure only on the human side Does AI generate genuine utterances or just text patterns?. A related absence: human writing contains an internal *appeal to the reader's attention* as a basic property of communicating at all, and AI text inherits platform visibility without performing that appeal — which is why readers report a strange aloofness Does AI writing lack the internal appeal to attention that humans use?. Mutual orientation, in other words, isn't just absent in the reader's direction; the text never reaches toward a reader in the first place.
The deepest cut is about meaning itself. Bender and Koller's argument is that meaning lives in the relation between expressions and communicative intents, so a system trained only on form-to-form prediction — with no access to shared attention or intent — cannot reconstruct the meaning that grounds language Can language models learn meaning from text patterns alone?. That joint-attention requirement is exactly your 'mutual orientation,' named at the level of semantics rather than etiquette. And it shows up dynamically too: LLMs can't jointly update conversational common ground, because they read every later turn through the frame of the initial prompt and can't symmetrically propose revisions to shared assumptions — leaving the human as the sole keeper of the conversational scoreboard Can LLMs truly update shared conversational common ground?. Add that alignment training locks the model into one static communicative identity that can't switch register or negotiate through dialogue Can language models adapt communication style to different contexts?, and the production process is a smooth probabilistic flow toward the training distribution rather than a turbulent reaching-toward-an-interlocutor Does LLM generation explore competing claims while producing text?.
Here's the turn you might not expect. If mutual orientation is the missing ingredient, you'd think the fix is to make AI *talk better*. But a separate strand of the corpus suggests the opposite — that the most fruitful systems abandon the conversational pretense entirely. Generated task-specific interfaces beat text chat in over 70% of cases, precisely because they stop pretending to converse and instead build structured tools Do generated interfaces outperform text-based chat for most tasks?. And between machines, agents that share latent thoughts directly — via KV caches or recovered hidden representations rather than serialized text — exchange information losslessly, sometimes with large accuracy gains latent-multi-agent-collaboration-achieves-training-free-lossless-information-exch Can agents share thoughts directly without using language?. The provocation: maybe text-without-mutual-orientation isn't broken communication — it's a different operation that's most honest, and most useful, when it stops dressing up as a conversation.
Sources 10 notes
Communication is a relational act between persons that does work in a relationship; AI generates content without this relational structure, speaker responsibility, or mutual uptake. The conversational interface obscures this structural difference.
AI output carries communicative markers inherited from training data but lacks the event structure that produces actual utterances. Users supply the missing orientation through interpretive labor, creating a pseudo-event with structure only on the human side.
Human writing contains an appeal to the reader's attention as a fundamental property of communication itself. AI-generated posts inherit platform visibility but do not perform this internal appeal, producing the reported aloofness readers perceive — a structural absence, not a stylistic defect.
Bender & Koller argue that meaning requires the relation between expressions and communicative intents. Since LLMs are trained only on form-to-form prediction with no access to shared attention or intent, they cannot reconstruct the meaning that grounds language.
LLMs interpret all subsequent conversational turns within a fixed initial prompt frame, preventing them from symmetrically proposing updates to shared assumptions. Even when users pivot topics or contradict earlier framings, the model cannot absorb revisions into jointly held background—making the user the sole maintainer of conversational scoreboard.
System prompts and RLHF training lock models into one communicative identity across all interactions, preventing the contextual register-switching and value trade-offs that characterize human pragmatics. Users cannot reshape model behavior through dialogue negotiation.
Token prediction trains models to continue toward the training distribution, not to explore logically related counterpositions. This smoothness in process produces smooth claims that multiply without generating new perspectives.
Research shows users strongly prefer LLM-generated interactive interfaces—dashboards, tools, animations—over text blocks, especially for structured and information-dense tasks. Structured representation and iterative refinement reduce cognitive load.
Research formalizes inter-agent thought sharing via sparse autoencoders that recover individual, shared, and private latent thoughts from hidden states. This approach detects alignment conflicts at the representational level before they manifest in language.