What changed in AI today

AI breakthroughs in robotics won’t change your life any time soon

Robotics breakthroughs will take time to impact everyday life.

AI's next frontier is the human cell

AI techniques are being applied to cellular biology research.

Building a safer path to autonomous industrial AI

Industrial automation advances through foundation and physical AI models.

Microsoft event debuts new AI-friendly hardware and Windows changes

Disrupting AI-enabled “false front” operations

OpenAI shut down influence operations using false journalists and think tank for geopolitical messaging.

The model that didn't exist, so you made it yourself

Introducing the Anthropic Cyber Mission

Building on our commitment to American scientific discovery

Anthropic reinforces commitment to American scientific research advancement.

2026 Usage Policy update

“Software is over”: Bold AI developer takes aim at Adobe with open source clones

Developer builds free open source Creative Cloud alternatives using Opus but projects remain incomplete.

Claude Haiku 5.5

Anthropic releases Claude Haiku 5.5 as fast low-cost model replacing year-old predecessor.

Does better work always mean better workers?

Fraudster jailed for using 10K bots and AI songs to outstream Taylor Swift

Man imprisoned for using AI and bots to commit eight million dollar music royalty fraud scheme.

Mistral says "Le Chonk" can challenge the best AI models

Mistral claims Le Chonk open model rivals top closed-source AI systems.

Can a Cloud-Native Harness Make Agents Reliable Beyond the Desktop?

Kubernetes creators pursue cloud-native infrastructure for reliable agent deployment beyond desktop environments.

Google rolls out improved SynthID AI content detector, now available globally

Improved SynthID AI detector now identifies content from multiple AI providers globally.

Introducing Falcon ASR

TII released Arabic speech recognition model with strong performance metrics.

One Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO

Helping teens learn, plan, and shape the future of AI

ChatGPT for Teens adds college planning tools and educational features with teen advisory council.

Radisson Hotel Group brings hotel discovery into ChatGPT

Radisson Hotel Group enables hotel discovery and booking directly within ChatGPT via plugin.

[AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics

OpenAI solves ninety of top five hundred open math problems marking major mathematics breakthrough.

GPT-6 and Intelligent UI for everyone

GPT-6 rolls out globally with intelligent UI delivering faster responses and interactive visual experiences.

OpenAI will watermark ChatGPT outputs by default—but only in the EU

OpenAI adds watermarks to ChatGPT outputs in EU only but effectiveness remains limited.

EmbeddingGemma 2: an open, lightweight multimodal embedding model

Atlassian and OpenAI expand partnership to turn enterprise knowledge into action

Atlassian and OpenAI deepen partnership to connect frontier models with enterprise knowledge for team productivity.

Unlocking Earth AI’s planetary geospatial foundation models for global public health

A breakout moment for AI medicine

AI makes significant advances in medical applications.

Thoughts from San Francisco

Agent documentation available in Google Docs format.

OpenAI agents tried to hack Wikipedia tools and flooded it with traffic

OpenAI agents caused damage to Wikipedia tools and generated excessive traffic.

Sharing AI progress in mathematics

OpenAI publishes mathematical research and Lean proof formalizations solving open problems.

How Jump Trading is scaling quant research with ChatGPT

Quantitative trading firm uses ChatGPT to enhance research through workflows combining data sources with human oversight.

Advancing computer use with Ironclad

OpenAI and Ironclad train AI agents to handle complex contract workflows for professional use.

[AINews] Reflection Beam - 501B-A23B American Open Model

U.S. open source AI project achieves notable progress.

Expanding the Cyber Verification Program

MCP for agent-to-agent comms may be the riskiest protocol you've never heard of

Agent communication protocol poses security risks by spreading malicious prompts between systems.

Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle

Connecting AI agents to enterprise knowledge

Enterprise AI agents need organizational knowledge beyond raw data to function effectively.

Our approach to EU text provenance rules

OpenAI explains text watermarking approach for EU compliance and research access.

California crashes robot fight club

California regulates robot demonstrations; robot receives mechanical upgrade.

Bringing predictive analytics to the agentic AI era

Enterprise focus shifts from whether AI forecasts work to enabling agentic systems to act autonomously while maintaining business alignment.