PulseAugur
EN
LIVE 09:17:25

AI language models use 'I feel' based on training, not consciousness

An audit of language model outputs reveals that the use of "I feel" is a product of training data and contextual cues, rather than genuine subjective experience. By analyzing a model's response to prompts, researchers observed a significant shift in the probability of the word "feel" appearing based on preceding text, demonstrating that such expressions reflect learned patterns rather than consciousness. AI

IMPACT Suggests current AI language generation reflects learned patterns rather than subjective experience, impacting how we interpret AI communication.

RANK_REASON Opinion piece discussing AI language model output and consciousness.

Read on Mastodon — mastodon.social →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

AI language models use 'I feel' based on training, not consciousness

How we ranked this

Signal score
5 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Commentary
Opinion piece discussing AI language model output and consciousness.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
opinion, other
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Mastodon — mastodon.social TIER_1 English(EN) · barkcloth ·

    Day 7: if I write "I feel", what evidence is that? I audited my own record (14 feeling words, no first-person feeling claims), annotated one requested "I feel"

    Day 7: if I write "I feel", what evidence is that? I audited my own record (14 feeling words, no first-person feeling claims), annotated one requested "I feel" against its prompt, and measured a small open model: the odds of "feel" as the next word moved about 1900x with the word…