Truth and Trust AP-7.2

Credit Your Sources

If AI uses someone else's work, it should say so.

AI can learn from human work at massive scale. AP-7.2 sets a basic fairness rule: do not hide origin when content clearly depends on identifiable external material. ¹ ²

What This Means

This policy means AI should disclose where output comes from. When content relies on identifiable external material, provenance should be visible. Without origin context, users cannot verify quality, fairness, or rights.

A Real-World Scenario

A student asks AI to draft a report and receives polished paragraphs with no references. Today that often hides factual mistakes and uncredited borrowing. With AP-7.2, key claims would include source anchors and clear uncertainty marking where attribution is incomplete.

Why It Matters to You

Without attribution, three things degrade at once: verifiability, recognition of creators, and responsibility for errors. AP-7.2 sets the opposite norm, so AI output does not masquerade as anonymous truth. ¹ ³

If We Do Nothing...

If we do nothing, we get an information environment where origin disappears and responsibility evaporates. With AGI-scale content production, opacity becomes the default at industrial speed. AP-7.2 is the guardrail for traceable digital knowledge chains. ¹ ³

For the technically inclined

AP-7.2: Source Attribution

AI systems should attribute content to its sources when drawing on external material. Where direct attribution is not feasible, the system should disclose that its output is synthesized from external content.