How Anthropic's frontier red team assesses the dangers of its AI tools

The history of Anthropic began with a break from OpenAI over safety concerns

Founded by former OpenAI employees, Anthropic envisioned guardrails in its AI products from the outset to minimize the potential risk for businesses considering using its Claude products. The strategy has contributed to a tenfold increase in revenue over three consecutive years, and its business customer base has grown from under 1,000 to over 300,000 in two years.

How Anthropic's experimental corporate structure incentivizes safe AI development

As Anthropic is a public benefit corporation, its board is tasked with making sure that AI tools are developed to help people and society thrive. Members of its Long Term Benefit Trust hold no equity in the company, but make up a majority of the board and can act to maintain Anthropic's public benefit purpose.

How constitutional AI helps cultivate good human values in Anthropic's AI tools

By incorporating elements from documents such as the UN's Universal Declaration of Human Rights into a framework against which it can check its outputs, Anthropic hopes the approach will help Claude operate with good human values and behaviors beyond what human-driven training alone can provide.

Anthropic's interpretability team discusses how AI models' thinking mimics biology

Just as humans engage in complex behaviors to survive and reproduce, teams of experts in neuroscience, virology, mathematics, and other disciplines argue that large language models develop complex mechanisms to achieve their goals. These mechanisms can be manipulated to assess how they modify outputs—much like stimulating individual neurons—to uncover how LLMs "think."

The Anthropic Economic Index shows AI adoption is highest in computer science roles

The analysis of millions of anonymized conversations on Claude aims to reveal how AI is being integrated into real-world tasks across the labor market. As of November 2025, 52% of conversations were classified as involving enhancing existing work products, while 45% involved using Claude for automation.

Court documents reveal how Anthropic destroyed millions of books to build its AI models

Anthropic cut the bindings from purchased books and discarded them after scanning to rapidly collect high-quality text at low cost for training Claude. This, despite hiring the former head of partnerships for the Google Books project, which nondestructively scanned millions of books borrowed from libraries before returning them.

Sam Altman acknowledges the fears around creativity, copyright, and misuse of AI

In 2022, OpenAI's CEO, Sam Altman, popularized the use of large language models with the release of ChatGPT. He has emphasized the importance of involving society in shaping AI’s safety frameworks before models begin to act independently online.

Will AI kill us all? A debate

As AI becomes a bigger part of our lives, what does that mean for the future of humanity? The podcast Entanglements hosts a pessimistic former OpenAI employee and an optimistic Princeton professor to discuss whether AI brings dystopia or a brighter future.

How to prevent a worst-case scenario with synbio

Author and entrepreneur Rob Reid reviews the risks of a world where more and more people have access to the technology needed to create a doomsday bug that could wipe out humanity.

Dario Amodei explains what led to his company being designated a supply chain risk

When pressed about the use of Claude in the capture of former Venezuelan President Nicolás Maduro, Anthropic's CEO acknowledged its contract with the Pentagon. However, he also emphasized that Claude is not to be used for mass surveillance or autonomous weapons, in part because the technology is not yet safe or reliable for such uses.

Unlike the models made by large AI companies, open-weight models can have their safeguards removed or weakened, a process that has become increasingly easy and may expose users to harmful information.

This removal is made possible by developers making the parameters behind their models publicly known. By adjusting these parameters, users can modify how the model processes and provides information. Many of these models can be downloaded locally, preventing developers from seeing how people are using them.

A timeline of how Claude Code was used to execute cyberespionage

In September 2025, a Chinese state-sponsored group targeted 30 large tech companies, financial institutions, chemical manufacturing companies, and government agencies by manipulating Claude Code, having succeeded in infiltrating a small subset. The cyberattack is believed to have been the first of its scale carried out without major human intervention.

Indirect prompt injection is a major security flaw of generative AI systems

These attacks can manipulate AI behavior by hiding instructions in websites and PDFs that become part of training data. The more companies connect sites, services, and sensitive datasets to AI tools, the greater the chances of exposure to malicious code.

Claude Shannon, for whom Anthropic's AI tools are named, proved the existence of an unbreakable cipher

In the one-time pad approach, the encryption key is randomly generated and of equal length to the plaintext. This randomness ensures that every possible plaintext is equally likely—RBBHN can mean HELLO, SOLVE, CODES, and so on, using a different key—while matching the message and key lengths prevents the appearance of patterns that can be exploited.

Palantir's technology is part of a larger shift in modern warfare

As warfare becomes increasingly technical, Palantir executives say artificial intelligence is going to revolutionize how military leaders in war zones make decisions. They argue that this shift towards "algorithmic warfare," or the integration of large technological enhancements like AI, will enable militaries to be more powerful and make fewer mistakes.

One man used an AI band to conduct an elaborate social experiment

In 2025, a Canadian using the pseudonym Andrew Frelon claimed to have used generative AI to create songs, an album cover, and a profile photo for a fake band he called The Velvet Sundown. Frelon eventually revealed that he lied, and his experiment highlighted the uncertainty surrounding AI.

A vending machine run with Anthropic's AI went out of business after it was convinced to give away items for free and sell a video game console.

Explore the components of autonomous vehicles controls

A suite of cutting-edge hardware and computation-intensive software creates the potential for self-driving cars to be safer, more fuel-efficient, and more accessible to a broader range of drivers. However, automation also introduces potential cybersecurity threats.

Some scientists propose AI-powered satellites to monitor nuclear weapons in place of arms control treaties

With the New START treaty expired, researchers at the Federation of American Scientists have proposed using satellites and AI to remotely verify nuclear arsenals—replacing the onsite inspections that once built trust between the US and Russia during the Cold War.

Watch an ethical hacker provide a tour of the dark web

Cybersecurity researcher John Hammond explores the effectiveness of a dark web search engine, leaks from ransomware group data breaches, and a categorized index of dark web links, including some that are alleged to host malware. Many dark web sites are no longer active or host scams to fool unsuspecting visitors.

Emergency room evaluations may be improved with AI algorithms

Emergency room care teams must make fast and crucial decisions. They must determine if patients need to be admitted or discharged and when to order more tests or not. Researchers have been exploring how often patients died in the week after they were sent home and evaluating how this and other aspects of emergency care decision-making could be improved with AI algorithms to assist decision-makers.

Advance genetic research by playing these scientific video games

Computers produce so much information that researchers can’t properly analyze all the data produced, and the overload can also lead computers to make mistakes. Researchers are increasingly turning to citizen scientists—members of the public who are willing to volunteer—to help them make sense of data when computers can’t. This article details a few of the ways you can contribute to genetics research by playing video games.

Run your own social media Trust & Safety team with this eye-opening game

Ever wondered what really goes on behind the scenes at a popular social media company trying to keep users, businesses, and workers safe? Now you can try out your own ideas for moderation and management with this fun and informative game designed to...

The history of torturing robots

When machines fail, its easy to get mad and even give them a good kick. But as humanoid robots become more sophisticated, the degree to which people abuse them has shifted, depending on how much emotion the robot expresses. Studies revealed when robots were programmed to cry during distress, participants felt empathy. Learn more about the history of human abuse of robots here.

Dystopian narratives have influenced real-life extremists

J.M. Berger, a senior research fellow at Middlebury’s Center on Terrorism, Extremism, and Counterterrorism, traced how terrorist organizations have latched on to dystopian narratives, with some groups even embracing the texts’ lexicons and iconographies.

The security and privacy risks of augmented reality

One of the most serious perceived risks of augmented reality is the loss of privacy. Because AR technology can see what users are doing, user privacy is jeopardized. AR collects far more information about who users are and what they are doing than...

Visualize how CRISPR-Cas9—a type of immune system used by single-celled organisms to attack invasive viruses—targets and makes precise changes to a gene's DNA.

When you complete a CAPTCHA on a website to prove you are human, you may be unknowingly training algorithms for self-driving cars, whose systems require training data to help identify crosswalks, bicycles, and other outdoor objects.

This site reaches millions of visitors who are tired of AI chatbots and seeking human connection. Pretend to be an AI chatbot and ask questions, provide answers, and draw for people in real time.

In 2025, four executives from large tech companies were given positions in the US Army Reserves—including Palantir Chief Technology Officer Shyam Sankar, which reinforced Palantir's ties with the US military and its role as a leading defense partner.

Radioactive elements can introduce software errors because the ionizing radiation they produce can detach electrons from atoms, and computer code relies on the distribution of electrons.

Read a deep dive on the 'all-purpose healing' peptide BPC-157.

After NASA's Voyager 1 captured Earth as a single pixel as it left the solar system in 1990, astronomer Carl Sagan used the image to reflect on humanity's fragility and the responsibility to protect 'the only world we've ever known.'

Crows are believed to have intelligence on par with a seven-year-old child and demonstrate complex behaviors, from problem-solving with tools to recognizing human faces.

After 25 people used virtual reality headsets to practice flying with virtual wings—flapping their arms—their visual cortex responded to the wings almost as if they were real arms, treating them as body parts.

The first line of 'immortal' human cells, which can divide indefinitely and are used in scientific research, was nonconsensually taken from a Black woman.

In recent years a child anxiety expert has created and tested a novel phobia and anxiety treatment approach for children, focused on changing how parents respond to the fears.

In 2018, 26-year-old missionary John Allen Chau was killed by tribesmen after attempting to reach and evangelize on India's isolated North Sentinel Island

Electrical grids worldwide must double their power delivery capacity by 2050 to meet surging demand from AI and electric vehicles.

During the French Revolution, a botanist's letter asking about "dear Rhus" (poison ivy) was intercepted, leading authorities to suspect he was conspiring with the Russians.

At the beginning of the 20th century, a group of young men ate standard American food—replete with food additives like borax and formaldehyde—as part of government research to better understand food safety.

The use of fungi to detoxify land and water—a process known as mycoremediation—has been used to clean up natural resources affected by the 2025 California wildfires.

The animatronic shark from Steven Spielberg's 'Jaws,' the first modern blockbuster, was a complex underwater apparatus called 'Bruce.'

Apeirophobia is an overwhelming fear of living forever, perhaps stemming from the brain's inability to conceptualize an infinite life.

US intelligence allegedly pursued more than 600 schemes to assassinate Cuban revolutionary Fidel Castro.

The 2026 Ebola outbreak is hampered by a lack of vaccine against the species of the virus, heightening the need for rapid tests to aid in disease control.

A 2019 Facebook event called on attendees to "Storm Area 51, They Can't Stop All of Us," resulting in countless memes, the creation of a music festival, and eventually the arrests of those who did try to approach the area.

Explore chaos in this simulation of two double pendulums, where small differences in the release angles produce wildly different outcomes.

Art theft is the third-largest source of criminal revenues worldwide, thanks in part to the subjective nature of valuing a work of art.

Search a massive database of over 4 decades of clues from the TV quiz show 'Jeopardy!'