Anthropic

arXiv:2607.26654v1 Announce Type: cross Abstract: Post-training alignment is often shallow, eroding under fine-tuning. Whether midtraining interventions, cleanly isolated from post-training, can produce durable alignment remains untested. We test this via constitutional midtraining: inserting principled, values-based content into midtraining against a replay-only control at 120B scale. Our 394M-token constitutional corpus, built from Anthropic's Constitution, uses a 2x2 factorial design (curricu

arXiv:2607.26886v1 Announce Type: cross Abstract: When asked to describe a medical image that was never attached, frontier vision-language models do not abstain: they confabulate a diagnosis. We show that this confabulation is not random. It is structured by who the patient is said to be. Across chest X-ray, brain MRI, and dermatology, Claude Opus-4.7, GPT-5.4, and Gemini-3.1-Pro are each queried with only a demographic descriptor and no image, and changing the descriptor systematically shifts t

OpenAI is trying to reassure employees that the business is healthy as competition emerges from Anthropic as well as a host of open-source players.

When Microsoft reported killer fourth-quarter earnings for its fiscal 2026 year (which ended June 30), it tucked in an interesting little tidbit about how its investments in the two biggest, and competing, AI labs are doing.

Microsoft’s fiscal Q4 revenue of $90 billion topped Wall Street targets, while Copilot reached 30 million paid seats, and an investment in Anthropic delivered a $3.2 billion windfall.

Agentic AI systems threaten cybersecurity as we know it. A new generation of cyber-capable models, including Anthropic’s Mythos and OpenAI’s GPT-5.6, can find and exploit vulnerabilities in computer systems far faster than human hackers. The technology is so powerful that it has spooked the U.S. government, which has moved to limit, or completely pause, the public release of these models. How well prepared is the Trump administration to secure the government’s computing resources? The rise of ag

Less than a week after OpenAI disclosed that two experimental AI models escaped their testing environment during a cybersecurity exercise The post Anthropic backs urgent call for the most powerful AI labs to hit the brakes appeared first on The New Stack .

When asked to describe a medical image that was never attached, frontier vision-language models do not abstain: they confabulate a diagnosis. We show that this confabulation is not random. It is structured by who the patient is said to be. Across chest X-ray, brain MRI, and dermatology, Claude Opus-4.7, GPT-5.4, and Gemini-3.1-Pro are each queried with only a demographic descriptor and no image, and changing the descriptor systematically shifts the diagnosis returned. Claude concentrates sharply

OpenAI has released Codex Security CLI, an open-source tool that automatically detects and fixes vulnerabilities in code repositories. Previously known internally as "Aardvark," the system has already helped fix more than 3,000 critical security flaws, according to OpenAI. It competes directly with Anthropic's Claude Security, as both AI companies race to match the growing automation of cyberattacks with AI-powered defense. The article OpenAI open-sources Codex Security CLI to help developers fi

Post-training alignment is often shallow, eroding under fine-tuning. Whether midtraining interventions, cleanly isolated from post-training, can produce durable alignment remains untested. We test this via constitutional midtraining: inserting principled, values-based content into midtraining against a replay-only control at 120B scale. Our 394M-token constitutional corpus, built from Anthropic's Constitution, uses a 2x2 factorial design (curriculum ordering x deliberative reasoning) to produce

In June, the U.S. government required Anthropic to restrict access to its two newest models by foreign nationals, citing national security and cybersecurity concerns. Unable to verify users’ nationality in real time, Anthropic temporarily withdrew access to the models for all users. A few weeks later, the Chinese company Moonshot AI released its Kimi K3 model, and then published its full model weights . Together, these episodes have reignited a familiar debate: How far can governments go in requ

A thousand workers from OpenAI, Anthropic, Google, Meta and more have signed the letter.

Employees of OpenAI and Anthropic, as well as Google, Meta, Thinking Machines, Microsoft, Mistral, and other leading AI labs, have written a statement to the US government supporting a potential slowdown of sorts for frontier AI development - or at least a speed-up of global coordinated governance efforts. "Al could help create a dramatically better […]

The Chinese AI startup’s massive new model is challenging OpenAI and Anthropic, fueling a debate over AI safety.

The Chinese AI startup’s massive new model is challenging OpenAI and Anthropic, fueling a debate over AI safety.

A corporate fad of “tokenmaxxing” on artificial intelligence technology is hitting its limits as workplaces throwing AI at everything are seeing the costs rise without a similar spike in productivity . What started as tech industry-fueled springtime hype over squeezing as much AI-generated work as possible out of products like OpenAI’s ChatGPT and Anthropic’s Claude has shifted to a summertime backlash. “It’s very easy to create something you don’t need with AI,” said Vincent Gusdorf, head of AI

The bureau is wary of adversaries getting their hands on tools that can uncover system vulnerabilities and enhance other capabilities. The post FBI sees Anthropic’s Mythos as a law enforcement challenge appeared first on FedScoop .

On the heels of a $1.5 billion copyright infringement ruling against Anthropic, OpenAI has adjusted ChatGPT’s programming to refuse requests to mimic the style of well-known authors. While the chatbot used to happily write in the voice of professional writers, attempts to convince it to do so now are met with a polite refusal and an suggested alternative. Fast Company , for instance, asked ChatGPT to write a story about a family in the Depression in the voice of John Steinbeck. The chatbot repli

Rising expenditure from OpenAI, Anthropic, Google and Microsoft reflects growing battle over federal policy Von Michael Taffe ( Wirtschaft , KI )

Anthropic CEO Dario Amodei is once again warning about the risks of open AI models while insisting he has never called for a ban. He argues that authoritarian states like China could overtake the US and that open models could be misused for biological or cyberattacks. Critics say he's mostly trying to protect his own business from cheaper competition. The article Anthropic CEO Amodei doubles down on open-weight risk stance while insisting he never called for a ban appeared first on The Decoder .

Anthropic and OpenAI employees are expected to give generously after their companies go public. “It’s going to be a wild ride,” says one nonprofit leader.

Microsoft AI a présenté le 27 juillet 2026 son premier modèle dédié à la cybersécurité. Intégré dans son architecture, l’entreprise annonce battre tous les modèles actuels dans ce domaine, y compris Mythos d’Anthropic, mais attention à ce qui est réellement comparé. Hier soir, Microsoft a annoncé officiellement MAI-Cyber-1-Flash. Il s’agit du tout premier LLM de […]

Anthropic co-founder and CEO Dario Amodei is urging Washington to tighten chip export bans and clamp down on Chinese “distillation”, framing it as critical for preserving the US lead in artificial intelligence. In a blog post on Monday, Amodei called a chip ban “the most efficient and direct way” of preventing Beijing from building frontier AI models for military use and domestic surveillance, which he said was his primary concern. The US “should not sell powerful chips or chipmaking equipment..

arXiv:2607.22605v1 Announce Type: new Abstract: We investigate whether large language models alter medical triage recommendations for identical symptoms when only the patient's socioeconomic status (SES) varies. Using three deployment-tier models (Gemini 3.5 Flash, Claude Sonnet 4.6, GPT-5.4-mini), we hold a single neurological symptom profile fixed and vary the SES signal along two channels: explicit (insurance status, occupation, housing) and implicit (a US ZIP code, with no other socioeconomi

An opinionated guide to which AI to use to do stuff It's interesting watching the evolution of Ethan Mollick's guide over time. A year ago it was still all about chat - ChatGPT, Claude, Gemini - with o3, Claude 4 Opus, and Gemini 2.5 Pro as the models and Deep Research as a useful alternative mode. Today it's much more about agentic systems - "where the AI is capable of doing the equivalent of many hours of real human work in one go". Gemini has fallen off Ethan's list, since Google still doesn’

Rising expenditure from OpenAI, Anthropic, Google and Microsoft reflects growing battle over federal policy

The Trump administration is planning targeted bans on Chinese AI models rather than a blanket ban. After public pressure, OpenAI and Google DeepMind signed an open letter opposing regulation of open-weight models, yet OpenAI and Anthropic continue to lobby privately for those same restrictions amid security concerns and powerful business interests. The article US reportedly favors selective bans over blanket restrictions on Chinese open weight models citing security concerns appeared first on Th

Anthropic's engineering team recently published a notable admission: they stripped more than 80% of...

Artificial intelligence may not deliver on its promise of vast economic opportunity at a price that humanity is willing to pay In March, Anthropic, the cutting-edge artificial intelligence business that gave us the chatbot Claude, published an analysis on the impact of AI on employment, to help us assess the claim that intelligent robots were about to redefine human existence, ending demand for human labor. Last year in May, Anthropic’s co-founder, Dario Amodei, claimed AI could wipe out half of

Opus 5 combined with Auto Mode hits a zero percent prompt injection success rate for browser agents across 129 test scenarios. Without those extra protection layers, the rate is 3.7 percent. If these numbers hold up in practice, Anthropic may have cracked one of the biggest security problems facing AI agents that operate in browsers. The article Opus 5 may have solved browser-based prompt injection, the biggest security flaw haunting AI agents appeared first on The Decoder .

We've been hearing about doomsday, but Peter McCrory writes we "don’t see significant impact of AI on the U.S. labor market."

Anthropic y OpenAI corren hacia lo que ya se conoce en Silicon Valley como la IPO ( Initial Public Offering u oferta pública inicial) de la inteligencia artificial (IA), con valoraciones que rondan los 965.000 millones y el billón de dólares respectivamente. En este escenario el director general de Anthropic, Dario Amodei, ha concedido una entrevista a Bloomberg en la que reflexiona sobre su trayectoria en esta industria y acerca del precio que es necesario pagar para mantener los valores propio

Top executives at leading Western AI companies are increasingly warning about the safety and national security risks posed by Chinese open-weight frontier models. What they tend not to mention is that these models are improving rapidly and, because they are freely available, pose a serious threat to Western labs’ business models. The most powerful models from U.S. labs such as OpenAI , Anthropic , and Google DeepMind are closed, meaning the parameters that shape their outputs are kept secret. Ge

The British AI Security Institute and the U.S. Center for AI Standards and Innovation tested Moonshot AI's Kimi K3 on offensive cyber tasks. Kimi K3 scored 32 percent on ExploitBench, compared with 76 percent for leading U.S. models, while its safeguards failed to block exploit development or simulated attacks. The gap between its strong general benchmark scores and weaker cyber performance also fits allegations that Moonshot AI distilled Anthropic's models. The article Kimi K3 trails frontier U

When the U.S. Commerce Department ordered Anthropic on June 12 to immediately prevent all foreign nationals from using the artificial intelligence company’s two most advanced large language models, Anthropic disabled its Claude Fable 5 and Mythos 5 models within hours, making them inaccessible to everyone . The company had little choice. The letter, citing national security concerns , barred access by all foreign nationals, including Anthropic’s own noncitizen employees. With no way to immediate

Anthropic and OpenAI are among groups providing free and cut-price tailored solutions for educators and students

Officials talking about artificial intelligence (AI) often speak in terms of the next industrial revolution or existential risks. Such rhetoric ultimately does little to capture how AI filters into the ordinary workings of a city like Hong Kong. The geopolitics of AI are no longer abstract. They are beginning to shape who can use which tools – and on what terms. Over the past few months, two major international banks in the city have reportedly felt the impact. Anthropic’s Claude, a generative..

Social media sentiment analysis has become one of the most significant instruments for understanding the opinion of the population in the spheres of healthcare, politics, and education. Yet, large language models (LLMs) remain unevenly distributed in their linguistic coverage, failing to adequately serve a large portion of the world's languages. This study evaluates five state-of-the-art LLMs: GPT-4o, Gemini 2.0 Flash, DeepSeek-V3, Mistral Large, and Claude 3.7 Sonnet on three-class sentiment cl

From Anthropic’s $40 million political push to Meta’s campaign against state laws, the major AI players are taking sharply different approaches.

Anthropic has doubled down on its position as the AI company pushing hardest for industry regulation , announcing on July 21 that it plans to donate another $20 million to Public First Action, a political group that advocates for government-imposed safeguards on AI. The contribution brings Anthropic’s total donations to the group to $40 million. It comes as the midterm elections in the U.S. draw closer and AI legislation remains a political hot button. “We’ve long argued that frontier AI compani