The AI Hate Progression

Reliability, Error, and Epistemic Concerns

• Many argue AI summaries and LLM outputs are substantially less reliable than prior search workflows, especially because:
  • They hallucinate confidently and unpredictably, rarely say “I don’t know”.
  • They remove source context and trust signals (domain, author, forum replies).
  • They can misread sources or even contradict them; some report summaries regularly misrepresent linked pages.
• Others counter that the internet was always full of misinformation; AI mainly changes convenience and scale, not the existence of errors.
• Several note that LLMs are fine for low‑stakes tasks or as dev tools, but dangerous when used for critical tasks (e.g., automotive torque specs, customer service decisions).

Consent, Copyright, and Data Use

• A dominant thread: AI training and deployment largely ignores user consent.
  • Creators feel they cannot post work, code, or writing without it being scraped for training.
  • There is anger that models are trained on copyrighted material and then compete with the originals.
• Some argue this is just an extension of long‑running consent erosion: search indexing, tracking, dark patterns, data breaches.
• Disagreement over “fair use”:
  • One side: training is transformative and thus fair use; reproduction is a separate issue.
  • Other side: this interpretation is a “miscarriage of justice” and contrary to copyright’s spirit.
  • Non‑US participants point out “fair use” is jurisdiction‑specific; US officials have also raised concerns about commercial model training.

Capitalism, Hype, and the AI Arms Race

• Many critiques target the AI “gold rush”: investor FOMO, oligopolistic control of compute, forced AI features in products, and AI‑washed layoffs.
• Debate over whether this is a failure of “capitalism” itself or of captured political systems and oligopolies.
• Some expect an AI bubble or “winter”; others see durable but narrower productivity gains.

Impact on Work, Creativity, and Everyday Life

• Creatives complain of platforms flooded with “AI slop,” harder discovery, and weakened livelihoods.
• Some workers use subsidized AI heavily at jobs, partly out of pragmatism, partly to exploit underpriced services.
• Call‑center examples show meaningful gains (multilingual triage), but others note many firms use AI to avoid hiring humans and degrade service.

Public Sentiment and Polarization

• Several note widespread distrust or dislike of “AI” as it is marketed (LLMs, image/video gen), even as people unknowingly rely on older ML features.
• Others emphasize AI products’ real popularity and argue the tech will normalize and recede into the background.
• Emotions range from mild annoyance to intense anger and boycott efforts; some commenters criticize both AI “zealotry” and anti‑AI “zealotry” as unproductive.