Jump to text
On this page
Copy and source
Published source confirmed at last check

Source changed 2026-10-14 19:58:00 UTC · snapshot created 2026-10-14 19:58:38 UTC · last check 2026-10-15 01:55:10 UTC

LDR.2:4.8 - Negotiate with or through capable AI

An AI may calculate alternatives, examine resource conflicts, formulate a counteroffer or accept bounded terms when an effective mandate permits it. Establish whose contribution it controls or represents, the scope of that mandate, the counterpart’s ability to identify and rely on the response, and the route for matters outside it. A convincing message from an agent is not evidence of a principal’s authorization.

Where qualified agents can exchange complete proposals and binding responses under agreed rules, a human negotiation meeting may be unnecessary. The substantive tests remain: feasible contributions, legitimate terms, consistent version and actual performance. Human preferences or consent need a permitted way of being represented; simulation of a person’s likely preference is not that person’s consent.

Compare the new arrangement’s setup, integration, computation, exceptions, oversight required by the work and consequences of error. Do not retain a human approval layer unless it supplies a necessary decision or guarantee. Do not remove a required approval merely because automated bargaining is cheap. C.11.DUA supports reconsidering both the arrangement and a disputed procedural burden through the proper authority.