Responsibility & Use of Artificial Intelligence

Responsibility & Use of Artificial Intelligence


The models, analyses and recommendations developed by Nobilys Group are intended to inform decisions relating to management, investment, resource allocation, risk steering and performance improvement.


They constitute decision-support tools. They do not replace the judgement of executives, internal validation procedures, or the legal, regulatory, social, financial or operational obligations specific to each organisation.


The outputs produced by an artificial intelligence model depend on the quality of the available data, the assumptions retained, the decision thresholds defined and the context of use. They may include uncertainties, estimation errors, false positives, false negatives or performance degradation over time, particularly in the event of changes in markets, behaviours, input data or operating conditions.


Any decision based on a model, score, forecast or recommendation must be subject to appropriate human review by the client. The client remains solely responsible for interpreting the results, for the decisions taken, for their implementation and for their consequences.


Nobilys Group does not guarantee that a model, score, forecast or recommendation will in all cases produce the expected economic, operational or financial outcome. Performance observed during a diagnostic, pilot or test depends on the available data, the assumptions retained, the context of use and the conditions of deployment.


Subject to the contractual provisions applicable to each engagement, Nobilys Group may not be held liable for decisions taken by the client, their implementation, or the consequences resulting from the use of models, scores, forecasts or recommendations outside the framework defined with Nobilys Group, without appropriate internal validation or without sufficient human review.


Where use cases involve personal data, automated decisions or systems likely to fall within a specific regulatory framework, the applicable obligations must be assessed on a case-by-case basis by the client with its legal, compliance or data protection advisers.


Nobilys Group supports its clients in structuring use cases, qualifying data, defining value criteria, governing models and setting the conditions for scaling. The specific terms of responsibility, use, validation and monitoring are defined in the contractual documents applicable to each engagement.