Researchers Warn Against Anthropomorphizing AI Model Intermediate Tokens as Reasoning
A position paper accepted at ICML 2026 argues that labeling intermediate token generation in language models as 'reasoning' or 'thinking' traces is a dangerous anthropomorphization. The paper contends that such terminology misleads users and researchers about the true nature of these outputs and how the models actually operate. The authors warn this framing leads to flawed research directions and misplaced trust in AI interpretability.
Researchers led by Subbarao Kambhampati have published a position paper, accepted at ICML 2026, challenging the widespread practice of describing intermediate token generation (ITG) in language models as 'reasoning traces' or 'thinking traces.' ITG refers to the text a model produces before arriving at a final answer, a technique that has become standard for improving performance on complex tasks. The paper argues that framing these outputs in human cognitive terms is not a harmless metaphor but a substantively misleading characterization that distorts understanding of how these models function. According to the authors, the anthropomorphization implies the intermediate tokens resemble human problem-solving steps and offer a transparent window into the model's cognition — claims they present evidence against. They call on the AI research community to adopt more neutral, technically accurate language when describing these outputs. The paper has gone through four revisions since its initial posting in April 2025, with the latest version substantially expanded to over 4 MB.
What's missing
The paper is a position paper rather than an empirical study, so the strength of its 'evidence' against anthropomorphization depends on arguments and cited work not detailed in the abstract. The specific harms or examples of 'questionable research' attributed to this anthropomorphization are not described in the available source material.
What different sources said
- arXiv cs.AICenter
Position: Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Traces!
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.