#safety

Specialization Beats Scale: A Strategic Variable Most AI Procurement Decisions Overlook

Specialization Beats Scale: A Strategic Variable Most AI Procurement Decisions Overlook When a model’s training history …

US scrambles to stop Internet users re-creating dead pilots’ voices

Pilots’ voices from the last seconds of a fatal cargo plane crash have been re-created by Internet sleuths using softwar…

Spotify is launching AI-generated remixes

Spotify and Universal Music Group (UMG) just announced a licensing deal that will allow users to prompt the creation of …

Tech researchers are suing the Trump administration over the future of online safety

Tech researchers are suing the Trump administration over the future of online safety In a lawsuit, the Coalition for Ind…

The Download: online safety’s future and climate tech’s big pivot

The Download: online safety’s future and climate tech’s big pivot Plus: SpaceX has filed for an IPO expected to be the l…

May 19, 2026 Announcements KPMG integrates Claude across its core business and workforce of more than 276,000 in strategic alliance

KPMG integrates Claude across its core business and workforce of more than 276,000 in strategic alliance KPMG—one of the…

May 19, 2026 Announcements Widening the conversation on frontier AI

Widening the conversation on frontier AI At Anthropic, we want to build AI systems that advance humanity and act for the…

Import AI 457: AI stuxnet; cursed Muon optimizer; and positive alignment

Import AI 457: AI stuxnet; cursed Muon optimizer; and positive alignment Welcome to Import AI, a newsletter about AI res…

May 18, 2026 Announcements Anthropic acquires Stainless

Anthropic acquires Stainless The frontier of AI is shifting from models that answer to agents that act—and agents are on…

Musk v. Altman week 3: Musk and Altman traded blows over each other’s credibility. Now the jury will pick a side.

Musk v. Altman week 3: Musk and Altman traded blows over each other’s credibility. Now the jury will pick a side. The tr…

Google updates its spam rules to include attempts to ‘manipulate’ AI

Google updated its spam policy to mark attempts to “manipulate” its AI model in search results as spam, including result…

Announcements PwC is deploying Claude to build technology, execute deals, and reinvent enterprise functions for clients

PwC is deploying Claude to build technology, execute deals, and reinvent enterprise functions for clients Anthropic and …

Behold, the Elon Musk jackass trophy

Yesterday, in Musk v. Altman, before the jurors came in, Sam Altman’s team passed up what looked — from a distance — lik…

Helping ChatGPT better recognize context in sensitive conversations

Helping ChatGPT better recognize context in sensitive conversations New safety updates help ChatGPT respond safely when …

Your doctor’s AI notetaker may be making things up, Ontario audit finds

In recent years, many overworked doctors have turned to so-called AI medical scribes to help automatically summarize pat…

May 14, 2026 Announcements Anthropic forms $200 million partnership with the Gates Foundation

Anthropic forms $200 million partnership with the Gates Foundation We’re partnering with the Gates Foundation to commit …

OpenAI Brings Its Ass to Court

Wednesday’s episode of the Musk v. Altman trial kicked off on Wednesday with a unique proposition: OpenAI wanted to brin…

Anthropic blames dystopian sci-fi for training AI models to act “evil”

Those with an interest in the concept of AI alignment (i.e., getting AIs to stick to human-authored ethical rules) may r…

Learning on the Shop floor

11th May 2026 - Link Blog Learning on the Shop floor. Tobias Lütke describes Shopify's internal coding agent tool, River…

I Work in Hollywood. Everyone Who Used to Make TV Is Now Secretly Training AI

My name on the platform is ri611. Or h924092b12ee797f, depending on who’s paying me. I work as an AI trainer. I assess w…

Introducing Trusted Contact in ChatGPT

Introducing Trusted Contact in ChatGPT Connecting you to someone you trust when it matters most. People use ChatGPT to l…

Trump Pivots on AI Regulation, Worker Ousted by DOGE Runs for Office, and Hantavirus Explained

This week on Uncanny Valley, the team discusses the surprising reports of the Trump administration seemingly reversing i…

ChatGPT’s ‘Trusted Contact’ will alert loved ones of safety concerns

OpenAI is launching an optional safety feature for ChatGPT that allows adult users to assign an emergency contact for me…

The balcony solar boom is coming to the US

The balcony solar boom is coming to the US Plug-in panels are getting popular—how do we make sure they’re safe? Dozens o…

Spooked by Mythos, Trump suddenly realized AI safety testing might be good

This week, the Trump administration back pedaled and signed agreements with Google DeepMind, Microsoft, and xAI to run g…

Mira Murati tells the court that she couldn’t trust Sam Altman’s words

Mira Murati, OpenAI’s former CTO, has testified under oath that CEO Sam Altman lied to her about the safety standards fo…

Advancing youth safety and wellbeing in EMEA

Advancing youth safety and wellbeing in EMEA Announcing our European Youth Safety Blueprint and EMEA Youth & Wellbeing G…

May 4, 2026 Announcements Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs

Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs Anthropic, Blacksto…

A new US phone network for Christians aims to block porn and gender-related content

A new US phone network for Christians aims to block porn and gender-related content Launching next week on T-Mobile's ne…

Elon Musk's 7 biggest stumbles on the stand at OpenAI trial

Elon Musk seems tired and cranky. On Thursday, he took the stand for the third day in a four-week trial stemming from hi…

Emergency First Responders Say Waymos Are Getting Worse

Emergency first-responder leaders told federal regulators in a private meeting last month that they were frustrated with…

Sam Altman is “the face of evil” for not reporting school shooter, says lawyer

OpenAI could have prevented one of the deadliest mass shootings in Canada’s history, a string of seven lawsuits filed We…

Our commitment to community safety

Our commitment to community safety Mass shootings, threats against public officials, bombing attempts, and attacks on co…

Google and Pentagon reportedly agree on deal for ‘any lawful’ use of AI

Google has signed a classified deal that allows the US Department of Defense to use its AI models for “any lawful govern…

Apr 28, 2026 Announcements Claude for Creative Work

Claude for Creative Work Creative professionals look to technology to expand what's possible in their work. Claude can't…

Apr 27, 2026 Announcements Anthropic names Theo Hourmouzis General Manager of Australia & New Zealand and officially opens Sydney office

Anthropic names Theo Hourmouzis General Manager of Australia & New Zealand and officially opens Sydney office Theo Hourm…

Man faces 5 years in prison for using AI to fake sighting of runaway wolf

A 40-year-old man was arrested after using artificial intelligence to generate a fake image of a runaway wolf that South…

GPT-5.5 Bio Bug Bounty

GPT‑5.5 Bio Bug Bounty Testing universal jailbreaks for biorisks in GPT‑5.5 As part of our ongoing efforts to strengthen…

Apr 24, 2026 Announcements An update on our election safeguards

An update on our election safeguards People around the world turn to Claude for information about political parties, can…

Import AI 454: Automating alignment research; safety study of a Chinese model; HiFloat4

Import AI 454: Automating alignment research; safety study of a Chinese model; HiFloat4 At what point do the financial m…

Apr 20, 2026 Announcements Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute

Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute We have signed a new agreement with Amazo…

Apr 14, 2026 Announcements Anthropic’s Long-Term Benefit Trust appoints Vas Narasimhan to Board of Directors

Anthropic’s Long-Term Benefit Trust appoints Vas Narasimhan to Board of Directors Vas Narasimhan has been appointed to A…

Responsible and safe use of AI

Responsible and safe use of AI Learn best practices for using ChatGPT safely and effectively. AI is a transformative new…

SOTA Normalization Performance with torch.compile

Introduction Normalization methods (LayerNorm/RMSNorm) are foundational in deep learning and are used to normalize value…

Introducing the Child Safety Blueprint

Introducing the Child Safety Blueprint A framework for combatting and preventing AI-enabled Child Sexual Exploitation Ch…

Announcing the OpenAI Safety Fellowship

Introducing the OpenAI Safety Fellowship A pilot program to support independent safety and alignment research and develo…

Apr 6, 2026 Announcements Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute

Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute We have signed …

Mar 31, 2026 Announcements Australian government and Anthropic sign MOU for AI safety and research

Australian government and Anthropic sign MOU for AI safety and research Today, Anthropic signed a Memorandum of Understa…

Introducing the OpenAI Safety Bug Bounty program

Today, OpenAI is launching a public Safety Bug Bounty(opens in a new window) program focused on identifying AI abuse and…

Inside our approach to the Model Spec

Inside our approach to the Model Spec As AI systems become more capable and widely used, we need a clear public framewor…

Helping developers build safer AI experiences for teens

Helping developers build safer AI experiences for teens Introducing a set of teen safety policies formatted as prompts f…

Building NVIDIA Nemotron 3 Agents for Reasoning, Multimodal RAG, Voice, and Safety

Agentic AI is an ecosystem where specialized models work together to handle planning, reasoning, retrieval, and safety g…

Creating with Sora Safely

Loading… The Sora 2 model and the Sora app offer state-of-the-art video generation with a new way to create together, an…

NVIDIA IGX Thor Powers Industrial, Medical, and Robotics Edge AI Applications

Industrial and medical systems are rapidly increasing the use of high-performance AI to improve worker productivity, hum…

How we monitor internal coding agents for misalignment

How we monitor internal coding agents for misalignment Using our most powerful models to detect and study misaligned beh…

OpenAI Japan announces Japan Teen Safety Blueprint to put teen safety first

OpenAI Japan announces Japan Teen Safety Blueprint to put teen safety first Strengthening age-appropriate protections, p…

vLLM Semantic Router v0.2 Athena: ClawOS, Model Refresh, and the System Brain Mar 10, 2026 · 23 min read Since v0.1 Iris, vLLM Semantic Router has made a large jump. In one release cycle, the project rebuilt its model stack, expanded routing into safety, semantic caching, memory, retrieval, and...

vLLM Semantic Router v0.2 Athena: ClawOS, Model Refresh, and the System Brain Since v0.1 Iris, vLLM Semantic Router has …

vLLM Semantic Router v0.2 Athena: ClawOS, Model Refresh, and the System Brain Mar 10, 2026 · 23 min read Since v0.1 Iris, vLLM Semantic Router has made a large jump. In one release cycle, the project rebuilt its model stack, expanded routing into safety, semantic caching, memory, retrieval, and...

vLLM Semantic Router v0.2 Athena: ClawOS, Model Refresh, and the System Brain Since v0.1 Iris, vLLM Semantic Router has …

Improving instruction hierarchy in frontier LLMs

Improving instruction hierarchy in frontier LLMs Introducing IH-Challenge, a training dataset that strengthens instructi…

Reasoning models struggle to control their chains of thought, and that’s good

Reasoning models struggle to control their chains of thought, and that’s good Why a limitation of frontier models is rea…

An update on our mental health-related work

An update on our mental health-related work Each week, more than 900 million people use ChatGPT to improve their daily l…

Advancing independent research on AI alignment

Advancing independent research on AI alignment We’re committing $7.5M to The Alignment Project to fund independent resea…

After Orthogonality: Virtue-Ethical Agency and AI Alignment

Preface This essay argues that rational people don’t have goals, and that rational AIs shouldn’t have goals. Human actio…

Bringing ChatGPT to GenAI.mil

Bringing ChatGPT to GenAI.mil Today, OpenAI for Government is announcing the next phase of our national security work: b…

Import AI 444: LLM societies; Huawei makes kernels with AI; ChipBench

Import AI 444: LLM societies; Huawei makes kernels with AI; ChipBench How can you quantify creativity? Welcome to Import…

EMEA Youth & Wellbeing Grant

EMEA Youth & Wellbeing Grant Supporting organizations improving youth safety and wellbeing in the age of AI. April 2026 …

The next chapter for AI in the EU

Key takeaways: - New program to train 20,000 SME across Europe with AI skills - €500,000 NGO grant to support research i…

Our approach to age prediction

We’re rolling out age prediction on ChatGPT consumer plans to help determine whether an account likely belongs to someon…

AprielGuard: A Guardrail for Safety and Adversarial Robustness in Modern LLM Systems

AprielGuard: A Guardrail for Safety and Adversarial Robustness in Modern LLM Systems In this work, we introduce AprielGu…

Updating our Model Spec with teen protections

We’re sharing an update to our Model Spec, the written set of rules, values, and behavioral expectations that guides how…

BNY builds “AI for everyone, everywhere” with OpenAI

BNY builds “AI for everyone, everywhere” with OpenAI With frontier capabilities from OpenAI, BNY enables employees to bu…

The Walt Disney Company and OpenAI reach landmark agreement to bring beloved characters to Sora

The Walt Disney Company and OpenAI reach landmark agreement to bring beloved characters from across Disney’s brands to S…

Update to GPT-5 System Card: GPT-5.2

GPT‑5.2 is the latest model family in the GPT‑5 series, and explained in our blog. The comprehensive safety mitigation a…

Import AI 437: Co-improving AI; RL dreams; AI labels might be annoying

Import AI 437: Co-improving AI; RL dreams; AI labels might be annoying Do you believe the singularity is nigh? Welcome t…

Stop Saying Boredom is Good for Kids

Chronic boredom is harmful to adults, causing stress, disengagement, and poor well-being. Academic researchers have show…

Funding grants for new research into AI and mental health

Funding grants for new research into AI and mental health Introducing a new program to award up to $2 million to support…

GPT-5.1-Codex-Max System Card

GPT‑5.1‑Codex‑Max is our new frontier agentic coding model. It is built on an update to our foundational reasoning model…

Strengthening our safety ecosystem with external testing

Strengthening our safety ecosystem with external testing Our approach to third party assessments for frontier AI. At Ope…

GPT-5.1 Instant and GPT-5.1 Thinking System Card Addendum

GPT‑5.1 Instant and GPT‑5.1 Thinking System Card Addendum As described in our blog, GPT‑5.1 Instant and GPT‑5.1 Thinking…

Introducing the Teen Safety Blueprint

Introducing the Teen Safety Blueprint A framework for building AI that protects, empowers, and creates safer experiences…

Introducing gpt-oss-safeguard

Today, we’re releasing a research preview of gpt-oss-safeguard, our open-weight reasoning models for safety classificati…

gpt-oss-safeguard technical report

gpt-oss-safeguard technical report Performance and baseline evaluations of gpt-oss-safeguard-120b and gpt-oss-safeguard-…

OpenAI gpt-oss-safeguard October 29, 2025 Ollama is partnering with OpenAI and ROOST (Robust Open Online Safety Tools) to bring the latest gpt-oss-safeguard reasoning models to users for safety classification tasks. gpt-oss-safeguard models are available in two sizes: 20B and 120B, and are permissively licensed under the Apache 2.0 license.

OpenAI gpt-oss-safeguard October 29, 2025 Ollama is partnering with OpenAI and ROOST (Robust Open Online Safety Tools) t…

Day Zero Support for OpenAI Open Safety Model

Day Zero Support for OpenAI Open Safety Model Fast and Affordable AI Inference For the World’s Latest Open Safety Model …

Addendum to GPT-5 System Card: Sensitive conversations

Addendum to GPT‑5 System Card: Sensitive conversations When we launched GPT‑5, we noted in the system card that we were …

Defining and evaluating political bias in LLMs

Defining and evaluating political bias in LLMs ChatGPT shouldn’t have political bias in any direction. People use ChatGP…

User Story Bilge Yücel DevRel Engineer Kelsey Sorrels Data Scientist at Telus AG How TAC Built an Agentic Chatbot with Haystack to Transform Trade Promotions Workflows See how TELUS Agriculture & Consumer Goods (TAC) gives users unprecedented access to their data with safety in mind October 6, 2025

How TAC Built an Agentic Chatbot with Haystack to Transform Trade Promotions Workflows See how TELUS Agriculture & Consu…

Launching Sora responsibly

Loading… Sora 2 and the Sora app combine cutting-edge video generation with a new way to create together, and we’ve made…

ENEOS Materials brings ChatGPT Enterprise to manufacturing

ENEOS Materials brings ChatGPT Enterprise to manufacturing Transforming the sector with AI-powered workforce solutions. …

Democratizing AI Safety with RiskRubric.ai

Democratizing AI Safety with RiskRubric.ai More than 500,000 models can be found on the Hugging Face hub, but it’s not a…

Detecting and reducing scheming in AI models

Detecting and reducing scheming in AI models Together with Apollo Research, we developed evaluations for hidden misalign…

Teen safety, freedom, and privacy

Some of our principles are in conflict, and we’d like to explain the decisions we are making around a case of tensions b…

A joint statement from OpenAI and Microsoft

A joint statement from OpenAI and Microsoft Loading… OpenAI and Microsoft have signed a non-binding memorandum of unders…

SafetyKit scales risk agents with OpenAI’s most capable models

SafetyKit scales risk agents with OpenAI’s most capable models From prototyping with early vision model previews to scal…

OpenAI and Greek Government launch ‘OpenAI for Greece’

OpenAI and Greek Government launch ‘OpenAI for Greece’ Today we’re launching ‘OpenAI for Greece’—a new partnership betwe…

Why language models hallucinate

At OpenAI, we’re working hard to make AI systems more useful and reliable. Even as language models become more capable, …

OpenAI and Anthropic share findings from a joint safety evaluation

Findings from a pilot Anthropic–OpenAI alignment evaluation exercise: OpenAI Safety Tests This summer, OpenAI and Anthro…

Collective alignment: public input on our Model Spec

Collective alignment: public input on our Model Spec We surveyed over 1,000 people worldwide on how our models should be…

OpenAI’s letter to Governor Newsom on harmonized regulation

OpenAI’s letter to Governor Newsom on harmonized regulation The US faces an increasingly urgent choice on AI: set clear …

From hard refusals to safe-completions: toward output-centric safety training

From hard refusals to safe-completions: toward output-centric safety training Introduced in GPT‑5, safe-completion is a …