Across small-group workplace problem‑solving formats that use external artifacts (whiteboards, shared docs, or AI assistants), does a generalized “artifact delay” principle—requiring some combination of individual unguided attempts and structured turn‑taking before any artifact is consulted—improve long‑term retention and equitable participation to a similar degree regardless of the specific artifact type, or do high‑agency artifacts like AI assistants require systematically different timing rules than low‑agency artifacts like static whiteboards to avoid reinforcing dominant‑solver patterns and social loafing?

Answer

A generalized artifact-delay principle (individual unguided attempts + some turn‑taking before consulting any shared artifact) is broadly beneficial for long‑term retention and equitable participation across artifact types—but high‑agency artifacts like AI assistants do require systematically stricter and more specific timing rules than low‑agency artifacts like static whiteboards or simple shared docs.

In practice:

• For low‑agency artifacts (whiteboards, basic shared docs), delaying artifact use until after everyone has attempted the problem and briefly shared their approach is usually sufficient to get most of the retention and equity benefits, assuming at least moderate psychological safety.
• For high‑agency artifacts (AI assistants that can propose complete solutions), similar delay rules are necessary but not sufficient. To avoid reinforcing dominant‑solver patterns and social loafing, you typically need:
  • Longer or more explicit unguided attempt windows before AI use,
  • Stronger gating on who can invoke the AI and for what (e.g., only to critique or extend human proposals),
  • And periodic AI‑free problems or phases so that the group practices full human solution-generation.

So there is a useful generalized delay principle, but as artifact agency increases, the timing rules must become more conservative and more tightly structured to prevent the AI from effectively becoming the de facto dominant solver that others loaf around.