Manifesto 2026 - AI Decision Sovereignty

Cited in institutional work

This manifesto is cited on pages 27 and 44 of the report of the French Senate's parliamentary mission on the alignment of artificial intelligence systems, delivered to the French minister for digital affairs in July 2026. Read the contribution and the cited passages →

The decision sovereignty is the capacity of an organisation to keep full control of its strategic, tactical and operational decisions, even while integrating artificial intelligence in its processes. It rests on one founding principle: decide before you integrate.

AI is transforming organisations quickly. Its capabilities are real and promising. That transformation must never mean the surrender of decisional responsibility . On the contrary, it demands that leaders be more vigilant, more strategic, more intentional than ever.

This manifesto is written for leaders, general management teams, and information systems and governance officers asking the essential questions: How to benefit from AI without ceding control to it? How to integrate these technologies while keeping effective governance ? How to build a doctrine for AI use that truly serves the interests of the organisation?

"AI never decides on its own. A human, named and accountable, must remain the decision-maker."

Manifesto 2026 - Principle II

The three principles of decision sovereignty

Decision sovereignty rests on three cardinal principles that structure every AI integration decision in an organisation.

Strategic clarity. Before activating an AI tool, the organisation must first clarify its strategic objectives, its values, its ethical and regulatory limits. This prior clarity is the non-negotiable condition.
Responsibility owned. AI never decides on its own. A human, named and accountable, must remain the decision-maker. AI is an assistance tool, never a replacement for judgement.
Control capacity. At any moment, the organisation must retain the technical, operational and legal capacity to take back control, to audit, to modify or to stop the use of an AI solution.

Why this doctrine now?

Organisations are under pressure. Technological competition is accelerating. AI's promises are seductive and often overstated. In this context, the risk is high of delegating critical decisions to poorly understood algorithms, integrated without real preparation, without explicit governance.

This risk is existential for organisations. An irresponsible AI integration decision can compromise reputation, legal compliance, operational continuity and the legitimacy of leadership.

This doctrine proposes a clear, actionable and resilient cadre to decide before you integrate. It is designed to be adapted to every organisational context, every sector, every maturity level.

Lines of action

The implementation of decision sovereignty is organised along five concrete lines.

Governance and responsibility. Establish clear roles, processes and decision bodies around AI.
Audit and transparency. Put in place mechanisms to audit AI systems, the outputs they produce, and their decisional impact.
Compliance and legal control. Ensure that AI use complies with regulation (GDPR, AI Act, sector-specific) and maintain the evidentiary capacity.
Training and awareness. Train decision-makers and operational teams to understand AI, its biases, its limits and its risks.
Operational resilience. Design AI integrations so that the organisation can operate without them in case of malfunction or incident.

This manifesto is not a brake on innovation. It is a responsible accelerator that lets organisations innovate quickly while keeping control of their strategic trajectories.

Download

Receive the Manifesto.

13-page PDF, delivered immediately. Your details are kept in our confidential contacts.

Discuss this publication ?

Fifteen to twenty minutes to discuss this subject applied to your organisation.

Get in touch