The agreement prohibits users from using Microsoft AI service outputs to train, create, or improve any AI technology, from attempting to reverse engineer AI models, and from using AI services for a list of nine prohibited applications including consequential automated decision-making without human oversight, emotional state inference from physical characteristics, facial recognition database creation through untargeted scraping, and social scoring or predictive profiling.
This analysis describes what Microsoft's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
Section 14.s establishes a set of prohibited AI use cases that mirror categories addressed in the EU AI Act, including prohibited practices related to biometric categorization, emotional inference, and real-time facial recognition. The prohibition on using AI outputs to train or improve other AI systems has direct operational implications for any user or organization processing Microsoft AI outputs in downstream model development workflows.
⚠ Use of Microsoft AI service outputs to train or improve other AI systems in violation of Section 14.s.iv may constitute a breach of the agreement, potentially resulting in account suspension or termination under the enforcement provisions of Section 4.b.iv.
Cross-platform context
See how other platforms handle AI Services Usage Restrictions and similar clauses.
Compare across platforms →"You may not use the AI services, or data from the AI services, to create, train, or improve (directly or indirectly) any AI technology. You may not use the AI services to discover any underlying components of the models, algorithms, and systems. For example, you may not try to determine and remove the weights of models or extract any parts of the AI services from your device. You agree that you will not use the AI services, including those that make decisions, or take actions, autonomously or with varying levels of human intervention: 1. To make decisions or take actions without appropriate human oversight that may have a consequential impact on any person's legal position, financial position, life opportunities, employment opportunities, or human rights, or may result in physical or psychological harm to a person or group of persons; 2. To deceive or intentionally misinform (e.g., false advertising), or deploy subliminal techniques...Excerpt from Microsoft's Services Agreement (Legacy)
1.
Enforcement risk, jurisdiction flags, contract triggers, and due diligence action items.
Get the research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean.
Section 14.s establishes a set of prohibited AI use cases that mirror categories addressed in the EU AI Act, including prohibited practices related to biometric categorization, emotional inference, and real-time facial recognition. The prohibition on using AI outputs to train or improve other AI systems has direct operational implications for any user or organization processing Microsoft AI outputs in downstream …
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Microsoft.