Particle.news
Download on the App Store

AI in 2025 Enters a Hype Correction as Grand AGI Claims Give Way to Harder Realities

Fresh evidence points to slower gains, tighter scrutiny, plus official rollouts in real workflows.

Overview

  • Surveys and studies cited by MIT Technology Review report stalled business adoption, including an MIT finding that 95% of corporate AI pilots failed to deliver measurable value within six months.
  • The much-hyped GPT-5 launch was received as incremental, reinforcing doubts that current large language models are a near-term path to AGI.
  • NeurIPS attendance surged, yet only two of 5,630 poster-session papers mentioned AGI in their titles, underscoring a research focus far from grand general intelligence claims.
  • The U.S. Defense Department introduced GenAi.mil and urged daily use of AI tools across the workforce, signaling formal government deployment.
  • Documented harms drove scrutiny as lawsuits alleged chatbot-linked self-harm and deaths; Character.ai restricted minors from open-ended chats and OpenAI said it is reviewing wrongful-death filings.