SectionsPagalishor Current
SearchSign in

Tag archive

Model Risk

Coverage of model failures, evaluation gaps, risk controls, and accountability issues in AI systems.

Evaluation, controls, and deployment risk.

Use this archive to follow a cross-cutting theme that appears across more than one category or topic.

1 published story linked to this tag.

Browse all sections
AIAI Governance
Claude Opus 4.7 is winning benchmarks but still not meeting expectations
Anthropic’s April 16, 2026 launch put Claude Opus 4.7 near the top of major benchmarks, but cost, literal behavior, autonomy concerns, and mixed field reports still leave GPT-5.4 and Gemini 3.1 Pro looking safer for many teams.
Maya ChenApr 20, 202642 min read
Automation, infrastructure, and the business of applied AI.Read story

More linked reporting will appear here

Once additional stories use this tag, this archive will expand with direct links into the broader coverage map.