Skip to content
Source · Daily Brief

AI Daily Brief — 28 December 2025

Sunday closed pre-Year-End cycle on retrospective + safety notes. OpenAI posts opening for new Head of Preparedness — executive role leading Preparedness team studying frontier-AI risks. CEO Sam Altman cited models “starting to present some real challenges,” specifically mental-health impacts and models “so good at computer security they are beginning to find critical vulnerabilities.” Euronews year-end essay: “2025 was the year AI slop went mainstream” — AI-generated low-quality content saturated feeds in 2025; mentions of “AI slop” across the internet rose 9× vs 2024, with negative sentiment peaking at 54% in October. Framed as the dominant 2025 culture story. NPR Weekend Edition Sunday: “The year in AI and culture” recapping AI/culture in 2025 — AI actress Tilly Norwood + music-label deals (UMG + Udio plan joint subscription service for 2026) + Disney’s $1B investment into OpenAI cited as marker events.

Top stories

  • OpenAI posts opening for new Head of Preparedness. Altman cites “real challenges”: mental-health impacts + models “so good at computer security they are beginning to find critical vulnerabilities.” via TechCrunch
  • Euronews: 2025 was year “AI slop went mainstream”. “AI slop” mentions across internet rose 9× vs 2024; negative sentiment peaked 54% October. Framed as dominant 2025 culture story. via Euronews
  • NPR Weekend Edition: “The year in AI and culture”. Tilly Norwood + UMG-Udio joint subscription for 2026 + Disney $1B OpenAI cited as marker events. via NPR

Who shipped

Nobody. OpenAI shipped a hiring page. Anthropic, Google, Meta, xAI, DeepSeek, Alibaba all dark.

Open-source pulse

No new drops. Chinese open-weight wave from Friday continues community uptake.

Money, infra & hardware

No new rounds. Year-end retrospective cycle dominates feeds.

Quiet corners

Altman’s “models finding critical vulnerabilities” framing is the substantive safety moment — first frontier-lab CEO explicit acknowledgment that current models can discover security vulnerabilities at consequential scale.

By the numbers

  • 0 — frontier-lab launches Sunday
  • 9× / 54% — “AI slop” mentions YoY growth / peak negative sentiment October
  • Most-mentioned topic: year-in-review + AI safety hiring
  • Quietest segment: everything

Compiled by AI Feed’s editor from verified web sources for 28 December 2025.