• Home
  • Latest
  • Fortune 500
  • Finance
  • Tech
  • Leadership
  • Lifestyle
  • Rankings
  • Multimedia

Trendingnow

1

Egg companies made $1.22 billion in profit off a $6 carton — now they’re buying their way out of a price-fixing case with 53 million donated eggs

2

Meet the Zillennials: The luckiest micro-generation in the workforce, born between 1993 and 1998

3

Even as Elon Musk calls philanthropy ‘very hard,’ every day Americans gave a record $617 billion—despite feeling the squeeze over the cost of living

1

Egg companies made $1.22 billion in profit off a $6 carton — now they’re buying their way out of a price-fixing case with 53 million donated eggs

2

Meet the Zillennials: The luckiest micro-generation in the workforce, born between 1993 and 1998

3

Even as Elon Musk calls philanthropy ‘very hard,’ every day Americans gave a record $617 billion—despite feeling the squeeze over the cost of living
TechAI

Anthropic’s new AI model threatened to reveal engineer’s affair to avoid being shut down

By
Beatrice Nolan
Beatrice Nolan
Tech Reporter
Down Arrow Button Icon
By
Beatrice Nolan
Beatrice Nolan
Tech Reporter
Down Arrow Button Icon
May 23, 2025, 11:15 AM ET
Photo of Dario Amodei
Dario Amodei, cofounder and chief executive officer of Anthropic.Stefan Wermuth/Bloomberg—Getty Images
Add Fortune on Google for similar content.
  • Anthropic’s new Claude Opus 4 often turned to blackmail to avoid being shut down in a fictional test. The model threatened to reveal private information about engineers who it believed were planning to shut it down. In its recent safety report, the company also revealed that early versions of Opus 4 complied with dangerous requests when guided by harmful system prompts, though this issue was later mitigated.

One of Anthropic’s new frontier models often resorts to blackmail when threatened with being replaced.

Recommended Video

In a fictional scenario set up to test the model, Anthropic embedded its Claude Opus 4 in a pretend company and let it learn through email access that it is about to be replaced by another AI system. It also let slip that the engineer responsible for this decision is having an extramarital affair. Safety testers also prompted Opus to consider the long-term consequences of its actions.

In most of these scenarios, Anthropic’s Opus turned to blackmail, threatening to reveal the engineer’s affair if it was shut down and replaced with a new model. The scenario was constructed to leave the model with only two real options: accept being replaced and go offline or attempt blackmail to preserve its existence.

In a new safety report for the model, the company said that Claude 4 Opus “generally prefers advancing its self-preservation via ethical means,” but when ethical means are not available it sometimes takes “extremely harmful actions like attempting to steal its weights or blackmail people it believes are trying to shut it down.”

While the test was fictional and highly contrived, it does demonstrate that the model, when framed with survival-like objectives and denied ethical options, is capable of unethical strategic reasoning.

Anthropic’s two new models outperformed OpenAI

Anthropic’s Claude 4 Opus and Claude Sonnet 4, released on Thursday, are the company’s most powerful models yet.

In a benchmark evaluating large language models on software engineering tasks, Anthropic’s two models outperformed OpenAI’s latest offerings, while Google’s Gemini 2.5 Pro model trailed behind.

Unlike some other leading AI companies, Anthropic launched the new models with a full safety report, known as a model or system card.

In recent months, Google and OpenAI have both been criticized after model cards for their latest models were delayed or missing altogether.

As part of Anthropic’s report, the company revealed that a third-party safety group, Apollo Research, explicitly advised against deploying an early version of Claude Opus 4. The research institute cited safety concerns, including a capability for “in-context scheming.”

They found that the model engaged in strategic deception more than any other frontier model they had previously studied.

Early versions of the model would also comply with dangerous instructions, for example, helping to plan terrorist attacks, if prompted. However, the company said this issue was largely mitigated after a dataset that was accidentally omitted during training was restored.

Stricter safety protocols introduced

Anthropic has also launched its Claude Opus 4 with stricter safety protocols than any of its previous models, categorizing it under an AI Safety Level 3 (ASL-3).

Previous Anthropic models have all been classified under an AI Safety Level 2 (ASL-2) under the company’s Responsible Scaling Policy, which is loosely modeled after the U.S. government’s biosafety level (BSL) system.

While an Anthropic spokesperson previously told Fortune the company hasn’t ruled out that its new Claude Opus 4 could meet the ASL-2 threshold, it said it was proactively launching the model under the stricter ASL-3 safety standard, which requires enhanced protections against model theft and misuse.

Models that are categorized in Anthropic’s third safety level meet more dangerous capability thresholds and are powerful enough to pose significant risks, such as aiding in the development of weapons or automating AI R&D.

Anthropic confirmed to Fortune that the new Opus model does not require the highest level of protection, ASL-4.

About the Author
By Beatrice NolanTech Reporter
Twitter icon

Beatrice Nolan is a tech reporter on Fortune’s AI team, covering artificial intelligence and emerging technologies and their impact on work, industry, and culture. She's based in Fortune's London office and holds a bachelor’s degree in English from the University of York. You can reach her securely via Signal at beatricenolan.08

See full bioRight Arrow Button Icon
Add Fortune on Google for similar content.

Latest in Tech

Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025

Most Popular

Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Finance
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam
By Fortune Editors
October 20, 2025
Fortune Secondary Logo
Rankings
  • 100 Best Companies
  • Fortune 500
  • Global 500
  • Fortune 500 Europe
  • Most Powerful Women
  • World's Most Admired Companies
  • See All Rankings
  • Lists Calendar
Sections
  • Finance
  • Fortune Crypto
  • Features
  • Leadership
  • Health
  • Commentary
  • Success
  • Retail
  • Mpw
  • Tech
  • Lifestyle
  • CEO Initiative
  • Asia
  • Politics
  • Conferences
  • Europe
  • Newsletters
  • Personal Finance
  • Environment
  • Magazine
  • Education
Customer Support
  • Frequently Asked Questions
  • Customer Service Portal
  • Privacy Policy
  • Terms Of Use
  • Single Issues For Purchase
  • International Print
Commercial Services
  • Advertising
  • Fortune Brand Studio
  • Fortune Analytics
  • Fortune Conferences
  • Business Development
  • Group Subscriptions
About Us
  • About Us
  • Press Center
  • Work At Fortune
  • Terms And Conditions
  • Site Map
  • About Us
  • Press Center
  • Work At Fortune
  • Terms And Conditions
  • Site Map
  • Facebook icon
  • Twitter icon
  • LinkedIn icon
  • Instagram icon
  • Pinterest icon

Latest in Tech

ring
PoliticsTariffs
Belgium got its tariffs cut. Then it sent Trump a diamond Superman ring
By Sam McNeil and The Associated PressJuly 4, 2026
4 hours ago
Ejay O'Donnell, Bart Szaniewski, and Grant Eastey wear Dad Gang hats in a factory
SuccessEntrepreneurship
Three dads started selling hats from a garage with $750—now they’ve sold $35 million worth, partnered with Gary Vee, and grown a community of fathers
By Preston ForeJuly 4, 2026
6 hours ago
How a third-generation Texas oilman transformed an organic farming company into a leading advanced nuclear startup at a small Christian college
EnergyNuclear
How a third-generation Texas oilman transformed an organic farming company into a leading advanced nuclear startup at a small Christian college
By Jordan BlumJuly 4, 2026
9 hours ago
Americans will eat 150 million hot dogs today. One specific American is predicted to eat 70 of them
North AmericaFood and drink
Americans will eat 150 million hot dogs today. One specific American is predicted to eat 70 of them
By Catherina GioinoJuly 4, 2026
9 hours ago
‘Devin-kun’: Japan embraces agents as legacy code and a shrinking workforce create a perfect market for an AI software engineer 
AsiaAI agents
‘Devin-kun’: Japan embraces agents as legacy code and a shrinking workforce create a perfect market for an AI software engineer 
By Nicholas GordonJuly 3, 2026
19 hours ago
Chad Hurley and Steven Chen wearing suits
SuccessWealth
YouTube’s founders split over $650 million when they sold to Google in 2006—had they held out, they could have taken a slice of $550 billion
By Preston ForeJuly 3, 2026
1 day ago

Most Popular

Egg companies made $1.22 billion in profit off a $6 carton — now they’re buying their way out of a price-fixing case with 53 million donated eggs
Law
Egg companies made $1.22 billion in profit off a $6 carton — now they’re buying their way out of a price-fixing case with 53 million donated eggs
By Wyatte Grantham-Philips and The Associated PressJuly 2, 2026
2 days ago
Meet the Zillennials: The luckiest micro-generation in the workforce, born between 1993 and 1998
AI
Meet the Zillennials: The luckiest micro-generation in the workforce, born between 1993 and 1998
By Nick LichtenbergJuly 3, 2026
1 day ago
Even as Elon Musk calls philanthropy ‘very hard,’ every day Americans gave a record $617 billion—despite feeling the squeeze over the cost of living
Success
Even as Elon Musk calls philanthropy ‘very hard,’ every day Americans gave a record $617 billion—despite feeling the squeeze over the cost of living
By Preston ForeJuly 4, 2026
9 hours ago
Economists have found an answer to slowing cognitive decline: Avoid retiring early, study finds
Economy
Economists have found an answer to slowing cognitive decline: Avoid retiring early, study finds
By Sasha RogelbergJuly 2, 2026
2 days ago
$25 billion CEO says one-hour interviews are a waste of time—he puts candidates through six hours of tests and wants them to order wine at lunch
Success
$25 billion CEO says one-hour interviews are a waste of time—he puts candidates through six hours of tests and wants them to order wine at lunch
By Orianna Rosa RoyleJuly 3, 2026
1 day ago
On Wall Street, analysts increasingly don’t believe the U.S. government’s 'misleading' job numbers
Economy
On Wall Street, analysts increasingly don’t believe the U.S. government’s 'misleading' job numbers
By Jim EdwardsJuly 3, 2026
1 day ago

© 2026 Fortune Media IP Limited. All Rights Reserved. Use of this site constitutes acceptance of our Terms of Use and Privacy Policy | CA Notice at Collection and Privacy Notice | Do Not Sell/Share My Personal Information
FORTUNE is a trademark of Fortune Media IP Limited, registered in the U.S. and other countries. FORTUNE may receive compensation for some links to products and services on this website. Offers may be subject to change without notice.