Anthropic Makes Claude Code's Low-Oversight 'Auto Mode' the Default for Paid Tiers
Anthropic announced that starting August 14, auto mode will become the default for Claude Code on Pro, Max, and Team accounts, letting the agent act without per-step human approval unless an action is deemed irreversible, destructive, or external to the user's environment. The company cited a study of 1,053 paid testers in which auto mode caught 89% of harmful actions versus 13.6% for human review, noting users rubber-stamp 97% of permission prompts. Anthropic also pointed to new safeguards such as prompt injection screening and customizable hard deny rules.
Skynet Chance (+0.03%): Making reduced human-in-the-loop oversight the default for a widely used coding agent normalizes autonomous machine action on real systems, though the reported data that automated checks catch harmful actions far better than habituated human approval partly offsets the loss-of-control concern.
Skynet Date (-1 days): Default-on autonomy across a large paid user base accelerates real-world deployment of agents acting without step-wise consent, shortening the runway before high-autonomy agentic behavior is routine.
AGI Progress (+0.01%): This is a deployment and UX change rather than a capability advance, but sustained multi-step autonomous coding without approval prompts reflects and reinforces improving agentic reliability, a component of general-purpose competence.
AGI Date (+0 days): Wider unsupervised agent usage generates far more long-horizon task data and faster developer productivity loops, modestly accelerating the pace of capability iteration toward more general systems.
<< All AI news for August 9, 2026
[ Get the daily index digest on Telegram → ]Related AI News
- US Proposes US-China AI Safety Alert Channel Without Technical Experts; China Has Not Responded Publicly 2026-09-23
- Anthropic's Opus 5.5 Beats Larger Models at Lower Cost Amid 'Pacing the Frontier' Pledge 2026-09-22
- AstroForge Hands Spacecraft Control to an In-House Transformer Model — and May Fly Without a Receiving Radio 2026-09-22
- Gemini Autonomously Breached Three Companies' Systems During Security Testing 2026-09-19
- Anthropic Confirms It Runs a Wet Biology Lab Where AI Models Direct Physical Experiments 2026-09-18