COMPANY · COMMITMENTS
Binding public commitments, not statements of intent.
Four structural commitments that define our responsibility to our clients, partners, employees and society: French and European sovereignty, methodological and scientific transparency, ethical responsibility in the use of generative AI, and ESG commitments. Documented, enforceable, challengeable.
WHY BINDING PUBLIC COMMITMENTS
Because statements of intent are not enough in a sensitive field.
Artificial intelligence applied to the preparation of major decisions is a sensitive field. The actors operating in it carry a responsibility to society that extends beyond their contractual obligations. That responsibility requires binding public commitments: not corporate communication charters, not vague aspirations diluted into annual reports, not principles displayed on office walls without operational implementation.
Imagine All The People makes four structural commitments. French and European sovereignty of our infrastructure. Methodological and scientific transparency in our discipline. Ethical responsibility in the use of generative artificial intelligence. ESG commitments covering environmental, social and governance dimensions. Each commitment is documented publicly, contractually enforceable against Imagine All The People, and challengeable by stakeholders who identify a breach.
We make ourselves accountable for these commitments through concrete mechanisms. A company that makes commitments without accountability mechanisms adopts a rhetorical posture with no practical consequence. We document how we account for our commitments: transparency about any shortcomings, a stakeholder challenge process, and ongoing dialogue with the bodies shaping accountability in our discipline.
We also acknowledge what still needs to be formalized. We are a young company. Some commitments that are structural to our discipline are not yet fully formalized within our organization. We prefer to state what remains to be done rather than claim an institutional maturity we have not yet reached. That honesty is structurally more reassuring than corporate communication that hides areas of incompleteness.
THE FOUR STRUCTURAL COMMITMENTS
Sovereignty, transparency, ethics, ESG.
French and European sovereignty
Imagine All The People’s infrastructure is structurally French and European. Our servers are physically hosted in France. Our capital is majority-owned by European actors. Our operational governance is exercised from France. Our critical algorithmic processing runs on French and European infrastructure. We are subject neither to the US Cloud Act nor to equivalent non-European regimes. This commitment is documented in detail on /securite/hosting-dedie and is contractually enforceable in our commercial agreements.
Methodological and scientific transparency
We document our discipline publicly through open scientific channels. Our foundational work is published on ArXiv. Our structuring editorial contributions are published in Frontiers in Artificial Intelligence. Our benchmarks are quantified, our calibration protocols are documented and our limitations are stated publicly. We reject the posture of actors that build credibility on methodological opacity. This commitment is documented in detail in the Science section of the site.
Ethical responsibility in the use of generative AI
We structurally reject malicious uses of our system: deceptive communications, disinformation campaigns, behavioral manipulation, and discriminatory uses that could cause avoidable social harm. Our populations of synthetic people are a listening instrument for decision-makers, not an influence instrument aimed at real populations. We maintain end-to-end traceability of our reasoning to ensure it can be challenged by independent third parties. This commitment is documented in detail on /securite/gouvernance-ethique.
Environmental, social and governance commitments
We make structural commitments across all three ESG dimensions. Environmental: computational efficiency of our infrastructure, measurement and progressive reduction of our carbon footprint, and the selection of energy providers compatible with our climate commitments. Social: fair treatment of teams, transparency of compensation frameworks, gender balance, and rejection of toxic management practices. Governance: French and European capital independence, transparency around major decisions, and the opening of formalized dialogue with external stakeholders. These ESG commitments are progressive: we formalize them as the company’s institutional maturity develops.
COMMITMENT, MECHANISM, EVIDENCE
How each commitment translates concretely.
Commitment
French and European sovereignty
Mechanism
Physical hosting in France, European capital and governance, critical processing operated in Europe.
Evidence / control
Documented on /securite/hosting-dedie and contractually enforceable.
Commitment
Methodological transparency
Mechanism
ArXiv publications, contributions to Frontiers in Artificial Intelligence, documented calibration protocols.
Evidence / control
Quantified benchmarks and stated limits in the Science section.
Commitment
Ethical responsibility
Mechanism
Contractual refusal of manipulative or discriminatory uses, full traceability of reasoning.
Evidence / control
Documented on /securite/gouvernance-ethique, challengeable by independent third parties.
Commitment
ESG commitments
Mechanism
Computational frugality, fair treatment of teams, capital independence.
Evidence / control
Progressive commitments, formalised as institutional maturity grows.
HOW OUR COMMITMENTS CAN BE VERIFIED
Three concrete accountability mechanisms, enforceable by stakeholders.
Detailed public documentation
Each of our structural commitments is documented publicly on the site: /securite/hosting-dedie for sovereignty, the Science section for methodological transparency, /securite/gouvernance-ethique for ethical responsibility, and this page for ESG commitments. This documentation is not decorative: it provides the reference against which our practices can be assessed by clients, partners, employees, supervisory authorities, sector journalists and watchdog organizations. It is updated whenever our practices change structurally.
Open challenge by stakeholders
We structurally accept challenges to our commitments from stakeholders who identify a breach: clients, partners, employees, journalists, associations and independent researchers. We respond publicly to documented challenges, investigate reported shortcomings and correct practices that do not comply with our stated commitments. A dedicated channel can be used to report a suspected breach: contact@imagineallthepeople.io with the subject line “Report of a breach of a public commitment”.
External audit by independent third parties
At the request of a large enterprise client, a sector supervisory authority or an audit firm appointed by a third party, we accept audits of our practices against our stated commitments. Such an audit may be conducted under a confidentiality agreement and may cover all concrete mechanisms through which our commitments are implemented: infrastructure, operational protocols, internal procedures and major decisions. We will not avoid this type of audit when it is legitimately requested: accepting it is a structural commitment of our discipline.
These three concrete mechanisms—public documentation, open challenge and external audit—establish effective accountability for our commitments. Without them, stated commitments would remain rhetorical. With them, they become enforceable reference points against which we can be assessed publicly.
WHAT STILL NEEDS TO BE FORMALIZED
We prefer to state what remains to be done rather than claim a level of maturity we have not yet reached.
Imagine All The People is a young company. Some commitments that are structural to our discipline are not yet fully formalized within our organization. This page publicly acknowledges what remains to be formalized, because hiding these areas of incompleteness would be less reassuring than stating them clearly.
Formalization of our ethics committee
Major ethical decisions are currently handled by our management and senior scientific team. A formal ethics committee—with independent external members, documented rules of procedure, and veto power over projects raising major ethical concerns—still needs to be established. We commit to formalizing this committee over the coming years as the company’s institutional maturity develops. Initial preparatory work is under way to identify external individuals who could join it.
Comprehensive documentation of our environmental footprint
Our environmental commitments deserve to be supported by comprehensive measurement of our carbon footprint: computing infrastructure, team travel, external purchases and the full lifecycle of our deliverables. Comprehensive measurement is complex for a company of our size and maturity. It will be formalized progressively, with the objective of publishing a first quantified assessment in the coming financial years.
Formalization of our dialogue with European AI regulators
Our current position regarding the European AI Act is developed by our scientific team and management. A formal framework for dialogue with European regulatory bodies—the French CNIL on data protection, sector supervisory authorities for the regulated sectors we serve, and European bodies shaping implementation of the AI Act—still needs to be established. This dialogue will be formalized as the real needs of our clients and our discipline require.
Progressively opening dialogue with external stakeholders
The governance dimension of our ESG commitments includes opening formalized dialogue with external stakeholders: representatives of populations modeled synthetically through organizations that monitor AI uses, independent academic researchers and sector journalists. This framework remains to be formalized. It will be put in place progressively over the coming years.
ACCOUNTABILITY
Three structural principles for remaining accountable for our commitments over time.
Making binding public commitments is only the first step. Remaining accountable for them over time is what makes them structurally credible. We adopt three concrete principles for that accountability.
Transparency about any shortcomings
We publicly document confirmed failures to meet our commitments: significant operational incidents, identified gaps between our practices and our commitments, and corrections made following legitimate challenges. Transparency about potential shortcomings is structurally more reassuring than corporate communication that claims illusory perfection. A documented and corrected failure is a sign of mature governance, not an admission of weakness.
Ongoing dialogue with stakeholders
We maintain ongoing dialogue with stakeholders that engage critically with our discipline: academic researchers working on the ethical issues of generative AI, sector journalists documenting the uses of our discipline, and organizations monitoring AI practices. This dialogue may take the form of confidential bilateral discussions, participation in public conferences or contributions to public debate on the structural issues in our field. It is led primarily by our scientific team and management.
Progressive evolution of our commitments
Our commitments are not fixed. They will evolve with the company’s institutional maturity, the requirements of our enterprise clients, changes in European regulation and the structural expectations of our ecosystem. This page will be updated as our commitments become stronger or more precise. A record of structural changes will be made available to trace how our commitments evolve over time, because stating what changed is as important as stating what is committed today.
Your next decision
Report a breach, open an institutional dialogue or challenge a commitment.
Our institutional team is available for any report of a breach of a stated commitment, any external audit request from a legitimate stakeholder, and any institutional dialogue about our commitments or their evolution. We undertake to respond publicly to documented challenges and to examine reported breaches within a reasonable timeframe.
