Here's your daily roundup of the most relevant AI and ML news for February 13, 2026. Today's digest includes 2 security-focused stories. We're also covering 6 research developments. Click through to read the full articles from our curated sources.
Security & Safety
1. Show HN: TrustVector – Trust evaluations for AI models, agents, & MCP
We built TrustVector (trustvector.dev for website) because “which model/agent/tool should we trust?” keeps getting answered with vibes, marketing, or outdated benchmarks. And a lot of our enterprise customers kept asking about it.TrustVector is an open-source evaluation framework + public directo...
Source: Hacker News - ML Security | 1 hours ago
2. AI safety leader says 'world is in peril' and quits to study poetry
Article URL: https://www.bbc.com/news/articles/c62dlvdq3e3o Comments URL: https://news.ycombinator.com/item?id=47007877 Points: 80
Comments: 52
Source: Hacker News - ML Security | 2 hours ago
Research & Papers
3. Toward Reliable Tea Leaf Disease Diagnosis Using Deep Learning Model: Enhancing Robustness With Explainable AI and Adversarial Training
arXiv:2602.11239v1 Announce Type: cross Abstract: Tea is a valuable asset for the economy of Bangladesh. So, tea cultivation plays an important role to boost the economy. These valuable plants are vulnerable to various kinds of leaf infections which may cause less production and low quality. It ...
Source: arXiv - AI | 18 hours ago
4. Zero-Sacrifice Persistent-Robustness Adversarial Defense for Pre-Trained Encoders
arXiv:2602.11204v1 Announce Type: cross Abstract: The widespread use of publicly available pre-trained encoders from self-supervised learning (SSL) has exposed a critical vulnerability: their susceptibility to downstream-agnostic adversarial examples (DAEs), which are crafted without knowledge o...
Source: arXiv - AI | 18 hours ago
5. When AI Persuades: Adversarial Explanation Attacks on Human Trust in AI-Assisted Decision Making
arXiv:2602.04003v2 Announce Type: replace Abstract: Most adversarial threats in artificial intelligence target the computational behavior of models rather than the humans who rely on them. Yet modern AI systems increasingly operate within human decision loops, where users interpret and act on mo...
Source: arXiv - AI | 18 hours ago
6. Temporally Unified Adversarial Perturbations for Time Series Forecasting
arXiv:2602.11940v1 Announce Type: new Abstract: While deep learning models have achieved remarkable success in time series forecasting, their vulnerability to adversarial examples remains a critical security concern. However, existing attack methods in the forecasting field typically ignore the ...
Source: arXiv - Machine Learning | 18 hours ago
7. When Agents Disagree With Themselves: Measuring Behavioral Consistency in LLM-Based Agents
arXiv:2602.11619v1 Announce Type: new Abstract: Run the same LLM agent on the same task twice: do you get the same behavior? We find the answer is often no. In a study of 3,000 agent runs across three models (Llama 3.1 70B, GPT-4o, and Claude Sonnet 4.5) on HotpotQA, we observe that ReAct-style ...
Source: arXiv - AI | 18 hours ago
8. Charting Empirical Laws for LLM Fine-Tuning in Scientific Multi-Discipline Learning
arXiv:2602.11215v1 Announce Type: new Abstract: While large language models (LLMs) have achieved strong performance through fine-tuning within individual scientific domains, their learning dynamics in multi-disciplinary contexts remains poorly understood, despite the promise of improved generali...
Source: arXiv - Machine Learning | 18 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.