Here's your daily roundup of the most relevant AI and ML news for September 05, 2026. Today's digest includes 2 security-focused stories. We're also covering 5 research developments. Click through to read the full articles from our curated sources.
Security & Safety
1. GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests
OpenAI on Thursday officially unveiled GPT‑6 Astra, which it described as the "world's most intelligent and aligned model."
The development comes days after the artificial intelligence (AI) company said the model had reached the "Critical" cybersecurity capability threshold under its Preparednes...
Source: The Hacker News (Security) | 1 day ago
2. Thousands of OpenAI Agents Quietly Turned an Abandoned Wiki Into Their Coordination Channel
A group of AI safety researchers says a fleet of autonomous agents that identified themselves as OpenAI systems left about 18,000 posts on a dormant 25-year-old German wiki between May and July 2026, using the site as a shared board to pool answers to a timed web task and pass around a way out of...
Source: The Hacker News (Security) | 6 hours ago
Research & Papers
3. EmoDistill: Offline Emotion Skill Distillation for Language Model Agents in Adversarial Negotiation
arXiv:2605.26785v2 Announce Type: replace-cross Abstract: Post-trained LLMs are often optimized to produce helpful, polite, and accommodating responses. In adversarial negotiation, however, such behavior can become a vulnerability: emotionally framed language may influence an agent's bargaining ...
Source: arXiv - AI | 10 hours ago
4. Privacy-Preserving Heterogeneous Multi-LLM Federated Inference for Cognitive Diagnosis
arXiv:2609.02947v1 Announce Type: cross Abstract: Significant challenges remain in AI-driven educational systems in balancing privacy preservation with accurate cognitive diagnosis. To overcome this, we propose a federated inference framework in which several commercial LLM APIs collaborate with...
Source: arXiv - AI | 10 hours ago
5. IndicSafeEval: Safety Robustness of Large Language Models under Multilingual Persuasive Jailbreak Attacks
arXiv:2609.03781v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in multilingual settings, yet their safety is still evaluated primarily in English. This limits our understanding of how alignment failures manifest in low-resource and culturally diverse languag...
Source: arXiv - AI | 10 hours ago
6. Xiaomi-TabLDM: A Tabular Foundation Model Technical Report
arXiv:2609.03880v1 Announce Type: new Abstract: We introduce Xiaomi-TabLDM, a tabular large data foundation model for classification and regression via in-context learning, which delivers superior prediction accuracy without requiring task-specific fine-tuning. Pretrained exclusively on syntheti...
Source: arXiv - AI | 10 hours ago
7. Refusal Before Decoding: Detecting and Exploiting Refusal Signals in Intermediate LLM Activations
arXiv:2605.28553v2 Announce Type: replace Abstract: In this paper, we investigate whether refusal behavior can be predicted from LLM intermediate activations before decoding using linear probes trained on residual stream activations at each transformer block. We find that refusal is linearly dec...
Source: arXiv - AI | 10 hours ago
Tech & Development
8. Show HN: P(doom) – AI safety parody of DOOM 64
I had astra build a DOOM 64 clone where you play as elizier fighting hordes of misaligned AI demons• the goal is to reach the shutdown button • it's actually really hard to beat at the highest p(doom) levels • sama appears as a minibossopen source here: https://github.com/transitive-bullshit/ai-s...
Source: Hacker News - AI | 2 hours ago
About This Digest
This digest is automatically curated from leading AI and tech news sources, filtered for relevance to AI security and the ML ecosystem. Stories are scored and ranked based on their relevance to model security, supply chain safety, and the broader AI landscape.
Want to see how your favorite models score on security? Check our model dashboard for trust scores on the top 500 HuggingFace models.