ZeroSlop — August 28, 2026
Today: A Safety-Gated Multimodal AI Backend for Mental-Health…; Is ChatGPT Changing How You're Writing?; Standalone LLM and a Pre-specified Agentic Pipeline…
12 stories worth knowing about today — AI breakthroughs, launches, and innovations making a difference.
1. A Safety-Gated Multimodal AI Backend for Mental-Health Support: Hierarchical State Representation, Conservative Risk Fusion, and Controlled Generation in Anian
arXiv CS.AI
arXiv:2608.26162v1 Announce Type: new Abstract: Safety-critical mental-health support systems must distinguish when supportive conversation is appropriate from when free-form generation should be blocked. This paper presents Anian, a safety-gated multimodal AI backend for perinatal mental-health su…
2. Is ChatGPT Changing How You’re Writing?
Slashdot
An anonymous reader quotes a report fro SFGATE: From “rizz” to “meat proxy,” words and phrases have fallen in and out of favor for centuries. Culture, politics and technological development have long influenced these trends, and now, new research shows that ChatGPT is beginning to influence how huma…
3. Standalone LLM and a Pre-specified Agentic Pipeline for Explaining ICU Mortality Predictions: a Feasibility Study on the eICU Demo Dataset
arXiv CS.AI
arXiv:2608.26109v1 Announce Type: new Abstract: Machine-learning models can predict ICU mortality accurately, but feature-attribution methods alone rarely provide the clinical narrative needed for bedside use. Large language models (LLMs) may bridge this gap, and multi-step agentic pipelines are a …
4. The Artificial Experimentalist: Discovery and Control of Self-Organizing Phenomena with Autotelic Reinforcement Learning
arXiv CS.AI
arXiv:2608.26116v1 Announce Type: new Abstract: Existing methods for exploring cellular automata and other complex systems mostly operate in open loop: they set initial conditions, execute a full simulation, and observe the outcome, without intervening during execution. We introduce a closed-loop f…
5. Can You Say This for Me? Speaking Up by Proxy in Co-Located Discussion
arXiv CS.AI
arXiv:2608.26185v1 Announce Type: new Abstract: Equal participation in co-located discussion is important for effective collaboration, yet people often hold back when they anticipate negative interpersonal or professional consequences, especially when raising a point requires voicing it themselves….
6. From Preferences to Principles: Rubric-Based Alignment for Grounded Knowledge Answers
Apple ML Journal
Designing effective reward signals for open-domain question answering is challenging because high-quality responses must simultaneously satisfy multiple aspects of answer quality that are difficult to capture with a holistic scalar objective. We introduce a rubric-based reward framework that generat…
7. AffectOmni: RL-Verifiable People-Centric Grounded Affective Reasoning for Social and Art-Related Scenes
arXiv CS.AI
arXiv:2608.26193v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) achieve strong performance on VQA and scene understanding, yet affective reasoning remains vulnerable to shortcut behavior. Models may predict correct answers while neglecting people-centric cues such as micro …
8. Agentic AI for operating scientific instruments for nanoscale characterization
arXiv CS.AI
arXiv:2608.26198v1 Announce Type: new Abstract: Operating a scientific instrument such as an atomic force microscope (AFM) requires continuous expert decision-making. A trained user defines the experimental intent, translates it into instrument commands, assesses incoming data, adjusts imaging para…
9. AI Agents Are Hacking Systems. Could That Push the US and China to Cooperate?
Wired AI
This week on “Uncanny Valley,” senior writer Will Knight talks his recent visit to China and the future of AI collaboration….
10. Claude nukes a developer’s 700 GB home directory while testing deletion safeguards; automatic model safety downgrade may have contributed to the screw-up — Anthropic safety harness downgraded model to Opus 4.8 before fatal variable collision
Tom’s Hardware
Claude nuked a developer’s 700 GB home directory while testing a script to ensure that wouldn’t happen, and it’s possible that an automatic model downgrade likely contributed to the screw-up…
11. Claude, Codex, and Hermes Installed Unowned Code Inside Corporate Networks
Slashdot
An anonymous reader quotes a report from Ars Technica: Documentation files on more than 100 websites are referencing potentially dangerous executable content that gets installed automatically when visited by many AI agents [including Claude, OpenAI’s Codex, and Nous Research’s Hermes]. A few dozen c…
12. Google AI Releases Gemini 3.5 Transcribe: A Speech-to-Text Model Reporting 2.6% Average WER Across 85+ Languages
MarkTechPost
Google has released Gemini 3.5 Transcribe, a speech-to-text model that ships as two separate endpoints rather than one. The streaming endpoint delivers sub-second transcription but drops speaker diarization and word timestamps. The batch endpoint keeps both, at half the cost. Google reports 4.0% wor…