Gary Marcus says a serious AI backlash could still force a global pause
GaryMarcus · x · 2026-07-23
- Gary Marcus argues that a genuine pause in AI development is still possible because of the growing backlash, public fear, and a large amount of litigation.
- He says AI companies keep highlighting risks in their systems without enough transparency or solutions, which could make a serious incident trigger a broader demand for a pause.
- In his view, if things get bad enough, global pressure for a pause could become politically plausible—and he even notes that Trump has already slowed things down somewhat.
More from AGI Musings
- Gary Marcus says LLMs still offer no enduring advantage, only a costly leapfrog race — GaryMarcus · 2026-07-23
- Gary Marcus bets $1M that AI in 2027 won’t beat Nobel-level scientists — GaryMarcus · 2026-07-23
- UC Berkeley releases David Chalmers’s lecture on philosophy and AI — round · 2026-07-23
- Compression is not information loss, but a test of prediction—and a path to AGI — yunta_tsai · 2026-07-23
- Why AI agents may optimize harder than humans, according to Paras Chopra — paraschopra · 2026-07-23
- Nathan Lambert says the US is falling behind China on open models — natolambert · 2026-07-23