AI HOT (Curated Pool)May 29, 2026, 11:00 UTC+8

Strengthening Societal Resilience with Rosalind Biodefense

用 Rosalind Biodefense 增强社会韧性

OpenAI launched Rosalind Biodefense and provides trusted GPT-Rosalind access to vetted developers and U.S. government partners; the post does not disclose model parameters or pricing.

Why it matters: HKR-H/K/R all pass: OpenAI launched GPT-Rosalind access for vetted developers and US government partners. Missing parameters, pricing, and eval results keep it below a major capability release.

Read the original ↗Export Markdown

Take

OpenAI is moving bio-risk capability from “whether to release” to “who gets the keys.” GPT-Rosalind access goes to vetted developers, U.S. government partners, and allied partners, while the post gives no parameters, pricing, or benchmark numbers. The concrete hook is access governance, not model capability.

The strongest line is buried in the safety history: OpenAI says ChatGPT agent was treated as High Capability in biology under its Preparedness Framework in July 2025. That makes Rosalind Biodefense a policy container for a capability class OpenAI already labels as dangerous. I buy half of it. Gated access beats a normal API, but the post does not specify whitelist criteria, logging rules, refusal boundaries, or audit terms for government use. That leaves the “trusted access” claim doing too much work.

Source: AI HOT (Curated Pool) · openai.com
Published: May 29, 2026, 11:00 UTC+8 · May 29, 2026