By using this site, you agree to the Privacy Policy and Terms of Use.
Accept
Felly ViralFelly ViralFelly Viral
Notification Show More
Font ResizerAa
  • Home
  • Nigeria
    NigeriaShow More
    Terrorists kill top governor’s SSA during prayers in popular northern state
    November 16, 2026
    United States offers faster response on about 27 visa routes
    November 15, 2026
    Australia to raise citizenship fee for foreigners from specified date
    November 15, 2026
    Full list: 15 countries, territories where Yoruba people are indigenes worldwide
    November 15, 2026
    Inside the legal fallout and quiet life Yaya Mayweather is navigating away from the headlines
    November 15, 2026
  • Technology
    TechnologyShow More
    Apple’s new study looks to improve the effectiveness of AI health tools
    November 16, 2026
    Vulcan rocket fails during launch, dealing major blow to SpaceX competitors
    November 15, 2026
    Jensen Huang is discussing a return to Nvidia, FT reports
    November 15, 2026
    With Starship catch, SpaceX still shows it can dazzle the world
    November 15, 2026
    Amazon has a plan to index the world’s online shopping data
    November 13, 2026
  • Sports
  • World News
    World NewsShow More
    Australian MP who lost both legs in alleged car attack discharged from hospital
    November 16, 2026
    After a nationwide ground stop, US faces aviation infrastructure questions
    November 16, 2026
    Steering failure preceded New Orleans cruise ship collision, says NTSB
    November 15, 2026
    Police ban protests near 10 schools and five universities in Paris
    November 15, 2026
    Prince Harry praises war veterans on surprise visit to Ukraine
    November 15, 2026
  • Politics
    PoliticsShow More
    Houthis advance in Yemen as government forces move to defend Aden
    November 16, 2026
    Russia's mega-rich are already imagining a future after Putin
    November 16, 2026
    Kano governor’s wife divorce case takes new turn, details emerge
    November 16, 2026
    Burnham pledges early intervention and ‘benefits backstop’ to tackle youth jobs crisis
    November 16, 2026
    Thousands of US flights still cancelled or delayed two days after grounding of planes nationwide
    November 16, 2026
  • Science
  • Entertainment
    EntertainmentShow More
    Clothing factories shut in Bangladesh as workers strike over pay
    November 16, 2026
    My MIL called herself the “real bride” at our wedding — Then my groom spoke
    November 16, 2026
    Listen to this: Armand Hammer’s Snakes & Ladders
    November 15, 2026
    Vulcan rocket fails, destroying 40 Amazon Leo satellites
    November 15, 2026
    England look to level ODI series against Australia LIVE!
    November 15, 2026
  • Crypto
    CryptoShow More
    Institutional Asset Tokenization Hits $10 Billion Milestone as Global Banks Expand Blockchain Settlement
    November 16, 2026
    Global Regulators Finalize Transatlantic Stablecoin Standards as Institutional Settlement Volume Surges
    November 14, 2026
    Chainalysis says hackers stole credentials and tokens for all of its customers
    November 6, 2026
    Chainalysis says hackers stole credentials and tokens for all of its customers
    November 6, 2026
    Binance says hackers used fake job interviews and an internal AI tool to pull off the $6.3 billion heist
    November 3, 2026
  • About Us
  • Contact
Reading: OpenAI names the four things it wants safety assessors to test
Share
Font ResizerAa
Felly ViralFelly Viral
  • Home
  • Entertainment
  • Science
  • Technology
  • Business
  • Crypto
  • Health
  • Nigeria
  • Politics
  • Sports
  • World News
Search
  • Home
  • About Us
  • Contact Us
  • Categories
    • Nigeria
    • Technology
    • Entertainment
    • Science
    • Health
    • Sports
    • Politics
    • World News
    • Business
    • Crypto
Have an existing account? Sign In
Follow US
Felly Viral > Blog > Technology > OpenAI names the four things it wants safety assessors to test
Technology

OpenAI names the four things it wants safety assessors to test

admin
Last updated: September 22, 2026 7:20 pm
admin Published September 22, 2026
Share
SHARE
September 22, 2026 at 7:20 pmIn: Technology

Credit: Zac Wolff on Unsplash That access should enable assessors to challenge our assumptions, identify risks we may have missed, and reach their own conclusions about the effectiveness of our safeguards. OpenAI published that on Tuesday, in a document setting out what it wants outside safety assessors to do. Lama Ahmad wrote it. She leads the company’s work with external safety experts.

The company said on Tuesday that it would open its models to independent scrutiny earlier in development. This document sits underneath that. It names four areas for assessment, and seven principles for how the work should run. Two definitions the rest depends on A safety claim, in OpenAI’s definition, is a specific assertion about a model that bears on its safety.

Evidence can test it. The claim should name the risks and conditions it covers, along with its assumptions and limits. A safety case gathers those claims into one structured argument. It explains why the company manages the risks of a given activity adequately.

It also has to state its assumptions, its uncertainties and whatever risk remains. Sam Altman used the same term last week. He argued for pacing rather than stopping, under a federal framework built on safety cases. The four areas The first is independent assessment of the safety cases themselves.

That spans training, evaluation, internal deployment and external deployment. OpenAI expects several assessors to take different parts, by expertise. It asks whether the evidence holds up. It asks whether the team actually followed the conditions of a safety case.

And it asks whether training methods reward deception, reward hacking or circumventing restrictions. The second is the safeguard stack. OpenAI wants assessors working with what it calls grey box access. They would test whether safeguards survive jailbreaks, and whether they stop capability uplift in cyber and biological domains.

It also wants authorised testing under realistic operating conditions. How do agents interact with access controls, sandboxing, and detection and response systems? Which defences prevent, detect or contain harmful actions, and which ones fail? Three questions in that section point straight at the company.

Do its misalignment monitors have gaps that could lead to loss of control? Does monitoring run across training, evaluation and deployment in a way nobody can easily disable? And how reliable does chain-of-thought monitoring stay as models get more capable? The third area covers capability evaluations under the Preparedness Framework, which tracks chemical and biological risk, cybersecurity and AI self-improvement.

OpenAI asks whether it sets its thresholds correctly. It also asks whether the evaluations get refreshed once models start topping them. The fourth is independent investigation of misalignment incidents. OpenAI names its own Hugging Face incident as the example.

Investigators would need cyber forensics skills, alignment expertise, and the ability to analyse chains of thought at scale. What the principles ask of both sides Both sides should agree the scope first, then pre-register the claims before any assessment begins. OpenAI wants it stated openly whether a claim came from the company or from the assessor. Conclusions have to say what the assessment left out.

Assessors should get access proportionate to the claims, within legal, security and intellectual property limits. Where direct access is impractical, the document points to a designated company representative, or to privacy-preserving mechanisms. They should also explain their methods, their criteria and their uncertainties. Reports should separate direct findings from interpretation.

Where no standard exists yet, assessors have to justify the criteria they picked. On independence, assessors should disclose conflicts of interest. That covers financial incentives, relationships with developers, and prior involvement in the work under review. OpenAI suggests recusal or exclusion periods, and says compensation arrangements should not shape findings.

Findings themselves should be specific enough to act on. OpenAI asks assessors to identify particular gaps, with enough detail for the lab to close them. Where it fits, reports should also draw out lessons for the people building, deploying and defending against agents. The parts that favour the lab Three of the seven principles give the company something back.

Where assessors cannot meet security requirements in their own environments, or where the data is especially sensitive, OpenAI says work on company-managed devices or premises may be appropriate. Where appropriate, it says labs should get a reasonable period to remediate issues before publication. The document puts no length on that period. And on publication, labs may request redactions of sensitive information.

Assessors keep editorial independence. They may note where substantive redactions happened, and what those redactions did to the report. The context the document leaves out OpenAI says it has already given assessors visible chain of thought access. It also claims unprecedented levels of access to confidential data and internal deployments.

The Preparedness Framework carries the third priority area. OpenAI disbanded the team that ran it in August, moving the work into other groups. Money goes unmentioned. In California, Senate Bill 813 created a route for independent verification organisations, and who pays an assessor has trailed the model ever since.

The document also builds on two OpenAI publications from the past week. One is its call for standards on Monday, which argued that the United States should lead an international effort on frontier AI. The other is the misalignment reporting framework it published on 16 September. Who would actually do this OpenAI says it is in conversation with multiple third parties about proposals matching these areas.

It names none of them. No single assessor can or should cover the urgent frontier safety questions, the company says. It adds that it will move deliberately while the independent evaluation ecosystem grows. The assessments described here are launch-agnostic rather than tied to a release.

Some would run for weeks. Others would run for several months.

You Might Also Like

Taiwan’s chipmaking giant TSMC to invest $1B in Anthropic, OpenAI

Six years later, Sony revisits its legendary XM4 headphones

Intel says it’s selling every affordable laptop chip it can make

Steam can now show game achievements in separate ‘sets’

Vulcan rocket fails during launch, dealing major blow to SpaceX competitors

Share This Article
Facebook Twitter Email Print
Leave a comment

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

What's Hot

Badenoch accuses Burnham of kicking defence spending into the long grass

Zack Polanski says he'll run for former PM Starmer's seat in by-election

Tragedy averted as fan slips and falls in front of Burna Boy's ₦9bn Bugatti

Rewetting peatlands may curb warming-driven CO₂ emissions

Davido reacts to video of Burna Boy taking delivery of his ₦9billion customised Bugatti Chiron

Bridging interfacial water structure and reactivity in photocatalytic hydrogen evolution at TiO₂ interfaces

Categories

Business

427 Articles

Politics

1523 Articles
- Advertisement -
Ad image

Categories

  • Sports
  • Politics
  • Technology
  • Science
  • World News
  • Business
  • Entertainment
  • Nigeria
  • Health
  • Crypto

About US

Welcome to Felly Viral, your destination for the latest trends, breaking stories, engaging updates, and conversations shaping the world around us
Quick Link
  • Home
  • About Us
  • Contact
  • Member Login
  • Profile
  • My Bookmarks
  • Privacy Policy
  • Terms and Conditions
Top Categories
  • Entertainment
  • Sports
  • World News
  • Politics
  • Technology
  • Science
  • Business
  • Health
  • Crypto

Subscribe US

Subscribe to our newsletter to get our newest articles instantly!

© Foxiz News Network. Ruby Design Company. All Rights Reserved.

Powered by CookieAdmin
Necessary cookies enable essential site features like secure log-ins and consent preference adjustments. They do not store personal data.
None
Functional cookies support features like content sharing on social media, collecting feedback, and enabling third-party tools.
None
Analytical cookies track visitor interactions, providing insights on metrics like visitor count, bounce rate, and traffic sources.
None
Advertisement cookies deliver personalized ads based on your previous visits and analyze the effectiveness of ad campaigns.
None
Unclassified cookies are cookies that we are in the process of classifying, together with the providers of individual cookies.
None
Powered by CookieAdmin
Welcome Back!

Sign in to your account