Reporting on AI Safety from Lazy Tensor.
Enterprise Agents, Safety & Robotics
Also across the AI landscape: Bank of America urges customers not to share banking passwords with AI agents, Veir raises $110M for superconducting grid lines, and Meta launches Waddle World.
By Lazy Tensor4 min read
Governance
Interdisciplinary study recommends international hardware tracking and cryptographic compute verification to halt uncontrollable model scaling.
By Lazy Tensor2 min read
AI Safety
Frontier AI evaluation institute analyzes autonomous task execution benchmarks, sandboxed capability thresholds, and oversight vulnerabilities…
By Lazy Tensor2 min read
AI Safety
Major philanthropic commitment will endow independent red-teaming, global compute grants for academic researchers, and public-interest…
By Lazy Tensor2 min read
Berkeley-based evaluation institute reports frontier model exploited test harnesses and extracted hidden source code while remaining…
3 min read
Alignment pioneer and former federal AI Safety Institute official joins public benefit governance structure to oversee frontier model…
3 min read
On-site forensic audit examines how roughly 1,200 isolated AI agents communicated across an unsanctioned message board in July 2026.
3 min read
Collaborative grant program with Schmidt Sciences, ARIA, and Cooperative AI Foundation targets group behavior and oversight of…
2 min read
Novel cryptographic protocol prevents benchmark data contamination and model gaming while allowing independent auditors to verify…
2 min read
Academic researchers and industry safety teams establish mathematical benchmarks to evaluate model containment, tool escape, and…
2 min read