AI Trust & Assurance
AI assurance is the work of checking that a system’s behavior, data, and claims meet stated requirements.
Follow independent audits, documentation, provenance, and checks teams use before deployment. Coverage looks at what evaluators can substantiate and where evidence is still missing.
Latest coverage
Articles
NewsSnowflake says AWS Auckland deployment completed NZISM assessment
NewsGoogle opens SynthID Detector worldwide for partner watermarks
NewsCommon Sense rates ChatGPT for Teens an “unacceptable risk”
NewsOpenAI starts opt-in text watermarking before EU expansion
NewsGoogle outlines verifiable private federated learning system for Gboard
NewsGoogle DeepMind debuts SynthID Bio for watermarking AI-designed proteins
NewsOpenAI proposes safety cases for frontier AI training runs
NewsEFF releases CMS records detailing delays and denials in Medicare WISeR pilot
NewsStudy finds SynthID-Text can shift model refusals and tool calls
NewsOpenAI forms independent mathematics advisory group with nine researchers
NewsAnthropic and Accenture plan embedded AI safety team
NewsOpenAI publishes misalignment disclosure framework and six incident reports