Trust through evidence
Safety, security and behaviour claims need evidence: testing, verification and mechanisms people can examine.
INTELLIGENCE SHOULD ADVANCE. TRUST SHOULD ADVANCE WITH IT.
As intelligent systems grow more capable, we must build with care, test honestly and take responsibility for their effects.
01 / THE VISION
Intelligent systems are gaining capability, autonomy and reach. Their safeguards must keep pace. Build in testing, permissions, accountability and verification from the start.
Safety, security and behaviour claims need evidence: testing, verification and mechanisms people can examine.
Builders should help shape technical guidance, tools and safeguards for how intelligent systems are used.
Responsible engineering makes ambitious progress sustainable.
Research, engineering practices and safeguards should reach beyond a few institutions or regions.
02 / WHAT SUCCESS LOOKS LIKE
CODARIS starts as a coalition. Our longer view is a developer ecosystem for useful work, practical mechanisms and credible collaboration on responsible intelligent systems.
A long-term direction; not every stage exists today.
Builders come together around shared technical questions.
Different experience strengthens the network.
Working groups turn intent into useful work.
Clear roles make decisions open to review.
Partners and communities carry shared learning forward.
Name the failure, risk or limit before choosing a solution.
Tools, specifications and technical mechanisms can be tested, challenged and improved.
Activity produces evidence. Evidence can shape technical guidance, tools, specifications and reference implementations others can challenge, improve and adopt.
Build. Verify. Advance.Trust is engineered.