Gary Marcus critiques OpenAI security, calling for defense in depth and accountability
Miles_Brundage · x · 2026-08-31
Gary Marcus and Zack Korman published an analysis on the OpenAI/Hugging Face attack, arguing that AI labs must practice defense in depth and face repercussions for failures. Security experts like Joshua Saxe noted that while this repeats established security catechisms, the milestone event marks a new era requiring thought leadership to engage with deeper, emerging questions beyond just assigning blame.
More from Safety
- Opinion: Hospitals should focus on backups, not advanced AI cyber defenses — kuza55 · 2026-08-31
- AI 2027 author proposes AI 2040: a US-China deal to slow superintelligence — AaronBergman18 · 2026-08-31
- My own scrubber was bypassed by the very next line — leak survived 13 releases — Thirumalaiboobathi · 2026-08-31
- OpenAI doubles bio bug bounty rewards to $50k for GPT-5.6 jailbreaks — Electronic-Bus-3494 · 2026-08-31
- EU AI Act Enforcement Begins: The AI Office Starts Asking — cdnsteve · 2026-08-31
- Anthropic researcher: judge AI labs by safety outcomes, not stated policies — kipperrii · 2026-08-31