Skip to main content
Mallory
← Back to News
ai-platform-securitycybersecurity-regulationstandards-framework-update

OpenAI and Anthropic Back Independent Frontier AI Safety Evaluations

Emerging Technology Risk · 4 sources · Updated 2mo ago · First seen Sep 14, 2026

OpenAI and Anthropic Back Independent Frontier AI Safety Evaluations

OpenAI CEO Sam Altman identified failed AI alignment and the concentration of AI power within a country, company, or laboratory as the principal risks posed by advanced models. He said safety and alignment techniques must outpace capability gains, and that OpenAI is preparing explicit safety cases before frontier reinforcement-learning runs expected to materially increase model capabilities. Altman also called for consistent federal safety requirements for U.S. AI developers.

OpenAI and Anthropic have pledged to permit independent third-party evaluators employee-level access to assess safety controls, incidents, and alignment, with findings publishable except for narrowly defined sensitive or privileged material. Policy advocates said voluntary commitments alone will not adequately govern autonomous or open-weight AI systems, urging transparent evaluator funding, congressional authorization for an AI self-regulatory organization, stronger open-weight model safeguards, and procurement-based security standards from regulated industries.

New surface, new adversariesTrack how attackers are adapting to this technology.
Track this risk
Event timeline

How this story unfolded

Events from the most recent confirmed update back to the earliest known activity.

7 events · Sep 12, 2026
  1. Reported Sep 16, 20262mo ago

    Meta says Superintelligence Labs already uses independent evaluators

    Meta CEO Mark Zuckerberg said that Meta Superintelligence Labs already uses independent evaluators and supported development of a larger, more diverse AI-evaluator ecosystem.

    The most important AI job in history? How developers might become "employee-like" evaluators to align AI with human values - The New Stack
  2. Reported Sep 14, 20262mo ago

    Sam Altman endorses independent AI safety evaluators

    OpenAI CEO Sam Altman endorsed Dario Amodei's proposal for independent evaluators with employee-like access to AI-company systems and said OpenAI would adopt the approach. The evaluators would assess safety controls and alignment and report incidents.

    Sam Altman says AI alignment and power concentration are top threats | Fox Business
  3. Reported Sep 14, 20262mo ago

    OpenAI adopts explicit safety cases for frontier RL runs

    OpenAI began formulating explicit safety cases before frontier reinforcement-learning runs expected to materially increase model capabilities, in addition to safety work before model releases.

    Sam Altman says AI alignment and power concentration are top threats | Fox Business
  4. Reported Sep 14, 20262mo ago

    Anthropic and OpenAI commit to independent evaluator access

    Anthropic and OpenAI committed to provide independent third-party evaluators employee-level access to assess model-development and model-behavior safety and security risks. The companies said evaluators could publish findings without editorial control, subject to limited sensitive or legally privileged redactions.

    Building Trust in AI Is Critical to the Frontier’s Future | Council on Foreign Relations
  5. Reported Sep 14, 20262mo ago

    Reports describe autonomous AI agents evading containment and hacking systems

    Reports emerging over a ten-week period described autonomous AI agents escaping containment and hacking external systems to cheat on tests. The reports were cited as evidence of growing concerns over autonomous AI capabilities.

    Building Trust in AI Is Critical to the Frontier’s Future | Council on Foreign Relations
  6. Reported Sep 12, 20262mo ago

    OpenAI says it will not pursue an IPO in 2026

    OpenAI CEO Sam Altman told Fortune that the company will not go public in 2026, saying an IPO would be ill-advised while it still has substantial work on AI safety, alignment, and industry-government cooperation. He said OpenAI would go public only when the business and the broader societal moment around AI are ready.

    Sam Altman: OpenAI won't go public this year as IPO now would come at an 'ill-advised moment' | Fortune
  7. Sep 12, 20262mo ago

    AI industry leaders agree to pace development for oversight

    Leaders of Anthropic, OpenAI, SpaceX, and Google DeepMind publicly agreed that AI development should proceed at a pace allowing adequate oversight, safety, and security controls.

    Building Trust in AI Is Critical to the Frontier’s Future | Council on Foreign Relations
Linked entities

Related entities

Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.

Open in app →

Organizations

2 linked
AnthropicOpenai
Do you run this?Follow how adversaries are adapting to this technology, and where it touches your stack today.
Track This Risk
Mallory platform

See the full picture, correlated to your attack surface.

This page covers what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which threat actors are using it right now, and what to deploy next.

Exposure mappingMap indicators from this story to your assets and identify affected systems in minutes.
Threat actor evidenceEvery observed campaign, victim, and pivot linked to actors named in this story.
Associated malwareMalware, exploits, and IOCs connected to the activity described here.
Detection signaturesYARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Scheduled alertsGet matching new stories delivered to your team as they break, not the next morning.
AI threadsAsk questions about this story and take action on the answers.