Engage stakeholders meaningfully, especially workers, their representatives and trade unions, affected communities and those most vulnerable to adverse impacts, as an integral part of every step: two-way, in good faith, responsive, with timely, truthful and complete information and a chance to give input before decisions that may affect them; build engagement into the AI system lifecycle (product purpose, design, dataset curation and validation, training and testing, use, post-deployment evaluation, multi-stakeholder and independent assessment, worker training where AI manages work); prioritise those most likely to be affected; plan modalities where product-level engagement is impractical; invest in stakeholder AI literacy; and respect workers' right to establish or join trade unions without interference.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.