Latest signals
AI, technology and society
From primary sources, research labs, regulators and trusted institutions.
Last updated 36 signals
Latest signals
The news stream
36 current signals
Oct 2, 2026 Announcements Anthropic invests $100 million to train 10,000 engineers and tackle the enterprise AI talent gap
Original sourceVictory! Court Rejects Government Effort to Dismiss Social Media Surveillance Lawsuit
Judge Allows Social Media Surveillance Lawsuit Against Trump Administration to Move Forward NEW YORK — A lawsuit filed by three labor unions against the Departments of State and Homeland Security for…
Original sourceHomomorphic Advantage Operator: Stabilizing Reinforcement Learning Under Fully Homomorphic Encryption Constraints
Privacy-preserving machine learning presents significant deployment challenges on the cloud for intelligent systems with confidential data. Fully Homomorphic Encryption (FHE) offers a compelling…
Original sourcePyPottery: an AI-powered end-to-end suite for pottery processing and publication
The study of ceramic materials constitutes a cornerstone of archaeological research, yet the post-production workflow for pottery documentation remains labor-intensive and creates significant…
Original sourceReconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents
Building reliable robot capabilities across diverse tasks requires substantial human effort to develop and maintain skills, design rewards, and integrate perception with control. We present…
Original sourceScholarCatalyst: A Benchmark for Retrieving Papers That Inspire New Research
What makes great scientists great? Even as AI systems start to make progress on open problems, scientists remain far ahead of them at sensing which prior idea, buried in an ever-growing archive of…
Original sourceWatch, Infer, Coordinate: Inferring Robot Partner Constraints for Zero-Shot Coordination
Robots operating in the physical world will increasingly need to coordinate with other robots, particularly in manipulation tasks where an object may be too large or heavy for a single robot to carry…
Original sourceWhere-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes
On-policy self-distillation has recently emerged as an effective approach for improving language-model reasoning by supervising students with a frozen or EMA version of themselves that receives…
Original sourceFaynt: Scaling and Optimizing Policies for Competitive Melee
We introduce Faynt, a family of 10M- and 75M-parameter Transformer policies for Super Smash Bros. Melee, each controlling all 26 characters with a single checkpoint. After reinforcement learning (RL),…
Original sourceGenerative modeling of intrinsically disordered protein regions by reinforcing sparse autoencoder features
Intrinsically disordered protein regions (IDRs) play central roles in cellular processes such as transcriptional regulation, signal transduction, and subcellular localization, yet their functional…
Original sourceSample complexity bounds for categorical Markov random fields via Discrete Diffusions
Many applications in statistics, economics, and physics require sampling from high-dimensional categorical distributions with local dependence structures. Examples include finite memory language…
Original sourceTACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning
Full-parameter fine-tuning of large language models (LLMs) incurs substantial optimizer state memory overhead, limiting the model sizes that fit on modern GPUs. Existing approaches either compress…
Original sourceTrust the Direction, Search the Step: Zero-and-First-Order Methods for LLM Fine-Tuning
Step-size selection remains a central challenge in large-scale neural network optimization; conservative steps slow convergence, while aggressive steps can destabilize it. We combine…
Original sourceWhen Do Intrinsic Rewards Lead to Exploration?
Intrinsic rewards are designed to guide exploration in reinforcement learning by assigning value to an agent's experience, for example through prediction error or learning progress. However,…
Original sourceCourt Agrees with EFF: Utah’s VPN Law Demands a Technical Impossibility
When state lawmakers attempt to rewrite how the internet works, users rely on courts to recognize that laws can’t make technical impossibilities a reality. That’s why we were happy to see that a court…
Original sourceHappy Opt Out October! Let’s Find Real Alternatives to the Tech Giants
Over the years, the major tech companies have found all sorts of ways to embed themselves into our lives. We often use their software, their AI tools, their social media, and their operating systems…
Original sourceNew tool lets users repair AI-generated 3D models, then fabricate them just the way they want
“InstructMesh” can generate designs for everyday objects that are easy to edit and fabricate for both experts and newcomers to 3D modeling.
Original sourceThe eternal complement
Advanced AI may matter most for the routine work behind breakthrough ideas. Explore why execution could shape the next economy and the pace of progress.
Original source📱 Hey Siri, How Do I Limit AI Data Access? | EFFector 38.17
With the launch of iOS 27, Apple is rolling out a variety of new AI features to its familiar voice assistant, Siri. But how are AI tools like these handling our data? In our latest EFFector newsletter…
Original sourceVictory! California Appeals Court Refuses to Revive Surveillance Tech CEO’s Meritless Lawsuit Against Journalist
When the rich and powerful try to use the court to silence negative reporting about themselves, it’s worth calling out that behavior for what it is: an attack on free speech. This is why EFF is happy…
Original sourceGemini 4 Argon: our next era of frontier intelligence
Original sourceIntroducing SynthID Bio
Proof of concept for watermarking AI-generated proteins while preserving biological function.
Original sourceMIT Transit Lab to develop an AI platform for public transit agencies
With $2.1 million funding from Google.org, the open-source Public Transit Intelligence Hub will unify public transit monitoring, operations, and passenger communication.
Original sourceThis game-playing AI is the new champ at Stratego
Able to defeat top-ranked human players and more efficient than other models, the new system could help decision-makers in military maneuvers or business negotiations.
Original sourceWhile the Country Rejects ALPR Mass Surveillance, SF Settles for Weak Safeguards
San Francisco's decision to retain its use of Automated License Plate Reader (ALPR) surveillance cameras belies what we know about this spying technology tool: it endangers residents and threatens the…
Original sourceThe effects of an “algorithmic monoculture” depend on the details
In a study focusing on hiring decisions, MIT researchers found the use of a single algorithm by many firms could benefit job seekers in certain situations.
Original sourceIntroducing dots
Dots by OpenAI are proactive assistants that can keep working across complex projects and everyday tasks. Learn how dots help you stay in control while work moves forward.
Original sourceIntroducing GPT-6.1 Sol
Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.
Original sourceIntroducing Claude Sonnet 5.5
A clear upgrade over Sonnet 5 that runs 30% faster and costs up to 30% less for most work.
Original sourceIntroducing the new Copilot with Home, Code and Autopilot
We’re reimagining Microsoft Copilot to enable work as it evolves and to help expand what every individual and every organization can accomplish in the flow of human ambition. Today we’re introducing…
Original sourceIntroducing Gemini 3.8 Live with Live Avatar
Original sourceAdvancing Private AI Compute with secure, server-side memory
Introducing private, server-side memory to Private AI Compute for personal AI.
Original sourceGrab and OpenAI bring practical AI skills to Southeast Asia
OpenAI and Grab launch GO Forward with AI, a regional programme helping 30,000 partners build practical AI skills across Southeast Asia.
Original sourceIntroducing MentalHealthBench
MentalHealthBench is an expert-informed benchmark for evaluating helpful and safe AI responses across realistic mental health conversations.
Original sourceIntroducing Claude Opus 5.5
Opus 5.5 performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5.
Original sourceIntroducing GPT-6 Sol and Luna
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
Original sourceSources we track
Sources we track
Primary sources, research labs, regulators and trusted institutions.
- OpenAI
- Anthropic
- Google DeepMind
- Microsoft
- MIT News
- Stanford HAI
- EFF
- EPIC
- European Commission
- NIST
- OECD AI
- NIH
- AI at Meta
- arXiv cs.AI
- arXiv cs.LG