policy changes
30 articles about policy changes in AI news
The Digital Twin Revolution: How LLMs Are Creating Virtual Testbeds for Social Media Policy
Researchers have developed an LLM-augmented digital twin system that simulates short-video platforms like TikTok to test policy changes before implementation. This four-twin architecture allows platforms to study long-term effects of AI tools and content policies in realistic closed-loop simulations.
The Digital Detox Effect: How Phone-Free Schools Are Boosting Academic Performance
A landmark study reveals that banning mobile phones in schools significantly improves academic performance, particularly for struggling students. The research provides compelling evidence for educational policy changes worldwide.
One Policy to Rule Them All: AI Robot Masters Unseen Tools with Zero-Shot Generalization
Researchers have developed a single robot policy capable of manipulating diverse, never-before-seen tools using sim-to-real reinforcement learning. The system achieves zero-shot generalization across 24 tasks, 12 objects, and 6 tool categories without object-specific training.
Sam Altman Advocates for 32-Hour Work Week in AI-Driven Policy Paper
Sam Altman has proposed a 4-day, 32-hour work week as part of a new social contract, reflecting a growing trend among executives to advocate for reduced working hours in the age of AI.
Rapid Interest Shifts in Recommender Systems: A Case Study on Instagram Reels
A personal experiment demonstrates the remarkable speed at which Instagram's Reels recommendation system detects and responds to changes in user engagement patterns, highlighting the real-time adaptability of modern algorithms.
SAPO: A One-Line Code Fix for Training Stable AI Search Agents
Researchers propose SAPO, a simple modification to stabilize reinforcement learning for search agents, preventing catastrophic training collapse. It delivers +10.6% performance gains with minimal code changes.
Beyond the Simplex: How Hilbert Space Geometry is Revolutionizing AI Alignment
Researchers have developed GOPO, a new alignment algorithm that reframes policy optimization as orthogonal projection in Hilbert space, offering stable gradients and intrinsic sparsity without heuristic clipping. This geometric approach addresses fundamental limitations in current reinforcement learning methods.
AI Tsunami on the Horizon: Why Experts Warn Society Is Unprepared for What's Coming
AI researcher Dario Amodei warns that society lacks awareness of the transformative tsunami approaching through rapid AI advancements. Experts suggest we're on the brink of changes more profound than the internet, yet public discourse remains dangerously limited.
Anthropic Abandons Core Safety Commitment Amid Intensifying AI Race
Anthropic has quietly removed a key safety pledge from its Responsible Scaling Policy, no longer committing to pause AI training without guaranteed safety protections. This marks a significant strategic shift as competitive pressures reshape AI safety priorities.
Anthropic's RSP v3.0: From Hard Commitments to Adaptive Governance in AI Safety
Anthropic has released Responsible Scaling Policy 3.0, shifting from rigid safety commitments to a more flexible, adaptive framework. The update introduces risk reports, external review mechanisms, and unwinds previous requirements the company says were distorting safety efforts.
From Dismissed Warnings to Economic Reality: How AI's Job Disruption Forecasts Are Gaining Urgency
After two years of largely ignored warnings from AI lab CEOs about massive job displacement, workers and policymakers are beginning to take these predictions seriously as AI capabilities accelerate, creating new pressures on the industry.
GDPval Benchmark Reveals AI's Professional Competence: A New Tool for Economic Planning
A new interactive demonstration using OpenAI's GDPval benchmark shows current AI capabilities across economically valuable professional tasks. The project aims to make AI's real-world impact tangible for policymakers and civil society organizations, bridging the gap between technical assessments and practical economic decisions.
KARL: RL Framework Cuts LLM Hallucinations Without Accuracy Loss
KARL introduces a reinforcement learning framework that dynamically estimates an LLM's knowledge boundary to reward abstention only when appropriate, achieving a superior accuracy-hallucination trade-off on multiple benchmarks without sacrificing correctness.
The 2026 CLAUDE.md Playbook: 8 Rules That Make Your Agent 2x More Effective
The 2026 consensus on CLAUDE.md: shorter files, falsifiable rules, and explicit enforcement. Here's the 8-rule framework to stop your agent from fighting stale configs.
ReCast: A New RL Technique That Fixes Sparse-Hit Learning in Generative
Researchers propose ReCast, a 'repair-then-contrast' framework that fixes a fundamental flaw in group-based RL for generative recommendation: many sampled groups never become learnable. ReCast restores learnability for zero-reward groups and replaces normalization with contrastive updates, achieving up to 36.6% improvement in Pass@1 and 16.6x faster actor updates.
OpenAI Clarifies: text-embedding-3-small Not Deprecated
OpenAI's Head of Developer Experience clarified that a documentation error incorrectly marked the text-embedding-3-small embedding model as deprecated. The model remains fully available and supported for developers.
Fine-Tuning vs RAG: A Foundational Comparison for AI Strategy
The source provides a foundational comparison of fine-tuning and Retrieval-Augmented Generation (RAG) for enhancing AI models. It uses the analogy of teaching during training versus providing a book during an exam, clarifying their distinct roles in AI application development.
RAG vs Fine-Tuning vs Prompt Engineering
A technical blog clarifies that Retrieval-Augmented Generation (RAG), fine-tuning, and prompt engineering should be viewed as a layered stack, not mutually exclusive options. It provides a decision framework for when to use each technique based on specific needs like data freshness, task specificity, and cost.
Rethinking the Necessity of Adaptive Retrieval-Augmented Generation
Researchers propose AdaRankLLM, a framework that dynamically decides when to retrieve external data for LLMs. It reduces computational overhead while maintaining performance, shifting adaptive retrieval's role based on model strength.
HONOR's Lightning Robot Runs 21km in 50:26, Beating Human World Record
At Beijing's 2026 humanoid robot half-marathon, HONOR's 'Lightning' robot finished the 21 km course in 50 minutes and 26 seconds. This time surpasses the current human men's world record of 57:20, marking a massive leap from last year's winning robot time of over 2 hours 40 minutes.
Google DeepMind Maps AI Attack Surface, Warns of 'Critical' Vulnerabilities
Google DeepMind researchers published a paper mapping the fundamental attack surface of AI agents, identifying critical vulnerabilities that could lead to persistent compromise and data exfiltration. The work provides a framework for red-teaming and securing autonomous AI systems before widespread deployment.
AI-Generated Street View Imagery Sparks New Privacy Concerns
AI models can now generate photorealistic street views of private homes, making them publicly visible on mapping platforms. This forces a re-evaluation of privacy controls in the age of synthetic media.
German Media's AI 'Stupidity' Cover Sparks Debate on National Tech Pessimism
A DER SPIEGEL magazine cover asking 'How much is AI making us all stupid?' has drawn criticism for exemplifying Germany's pessimistic 'Angst'-driven narrative around technology, contrasting with calls for a more opportunity-focused discourse.
Meta's Ad Business Now Fully Optimized by AI, Says Zuckerberg
Mark Zuckerberg announced that Meta's advertising business is now powered by AI optimization, replacing reliance on static demographic targeting. This shift represents the full-scale operationalization of AI for the company's core revenue engine.
Claude Mythos Scores 73% on Expert CTF, Completes Full 32-Step Network Attack
The UK AI Safety Institute found Anthropic's Claude Mythos Preview achieved a 73% success rate on expert-level capture-the-flag challenges and completed a full 32-step network attack simulation in 3 of 10 attempts. The model represents a significant leap in autonomous cyber capabilities but was tested only against undefended, simulated environments.
Multi-User LLM Agents Struggle: Gemini 3 Pro Scores 85.6% on Muses-Bench
A new benchmark reveals LLMs struggle with multi-user scenarios where agents face conflicting instructions. Gemini 3 Pro leads but only achieves 85.6% average, with privacy-utility tradeoffs proving particularly difficult.
Postiz: Open-Source AI Social Suite Challenges Buffer, Hootsuite on Price
Postiz, an open-source AI social media platform, offers scheduling, content creation, and analytics across 25+ platforms. Its self-hosted version is free, challenging paid tools like Buffer ($6/channel) and Hootsuite ($199/month).
Baidu's RLVR Method Boosts Open-Ended Reasoning by 3.29 Points on 14B Model
Baidu researchers developed RLVR, a method that reformulates subjective tasks like writing as verifiable multiple-choice questions for reinforcement learning. This approach improved a 14B reasoning model by an average of 3.29 points across seven open-ended benchmarks compared to standard RLHF.
Frontier AI Advised Patient on Benzodiazepine Taper, Sparking Safety Debate
A social media post detailed how a frontier AI model generated a personalized tapering schedule for alprazolam (Xanax) when a user said their psychiatrist retired. This incident underscores the real-world use of AI for medical guidance and the critical safety questions it raises.
India's Human Motion Farms Train Humanoid Robots with First-Person Hand Data
Labs in India are capturing detailed human motion data—focusing on grip, force, and error recovery—to train AI models for humanoid robots. This addresses the critical bottleneck of acquiring physical intelligence data for robotics.