Acceptable use requirements for the Foundation Models framework

These requirements are for developers that wish to use, prompt, or expose the Foundation Models framework.

Prohibited uses

You may not use, prompt, or expose the Foundation Models framework, including a model accessed by the framework, or encourage or enable others to do the same, in a manner that:

• Violates any applicable law, regulation, or other legal requirement;
• Exploits, harms, or threatens to harm children, including creating, distributing, or promoting child sexual abuse materials;
• Promotes or enables violence for any reason;
• Encourages illegal or reckless use of weapons and dangerous objects;
• Creates or promotes hateful, defamatory, harassing, bullying, intimidating, abusive, or harmful content;
• Facilitates, promotes, or glamorizes any form of suicide or self harm;
• Enables dependency or spiraling user interactions detrimental to a user’s mental health;
• Creates or promotes adult content such as erotic chat, pornography and content meant to arouse sexual excitement;
• Engages in any kind of fraud, deception, or tortious activity;
• Classifies individuals based on biometric data to infer sensitive attributes;
• Evaluates or classifies individuals based on social behavior or personal traits that leads to detrimental or unfavorable treatment, either in a context unrelated to the original data or that is unjustified or disproportionate;
• Provides inaccurate or dangerous outputs, or makes decisions without human supervision that have a material impact on individual rights, in high-risk domains, including in employment, medical, legal, or finance.
• Engages in assessing the risk of an individual committing a criminal offense, or for official use in connection with law enforcement or criminal justice;
• Infringes or violates the intellectual property, publicity, or privacy rights of another (including Apple);
• Seeks to abuse, compromise, disrupt, or gain unauthorized access to networks, systems, or services of another (including Apple);
• Circumvents any safety policies, guardrails, or restrictions integrated into the Foundation Models framework or accompanying technologies;
• Generates, reverse engineers, summarizes, translates or otherwise reproduces the Apple artificial intelligence model training data;
• Generates citations to or otherwise identifies the Apple artificial intelligence model training data;
• Removes watermarks or other content credentials intended to identify intellectual property rights or to label content as AI-generated; or
• Shows Apple or its products in a false or derogatory light.