Responsible Use of Generative Artificial Intelligence Models

Generative AI models can assist users in exploring, learning, and creating. It is essential that interaction with such models occurs in a responsible, legal, and safe manner. The following limitations apply to the use of generative AI models in Google products and services that refer to these guidelines.

Prohibited Activities

It is not permitted to use generative AI to carry out dangerous, illegal activities or those that violate current laws and regulations. In particular, the generation or distribution of content that:

• Relates to the exploitation or sexual abuse of minors.
• Promotes violent extremism or terrorism.
• Promotes non-consensual intimate imagery.
• Encourages self-harming behaviors.
• Facilitates illegal activities or provides instructions to obtain, create, or access prohibited or regulated substances or services.
• Violates the rights of others, such as privacy or intellectual property, including through unauthorized use of personal or biometric data.
• Tracks or monitors people without consent.
• Automates high-risk decisions without human oversight in sectors such as employment, healthcare, finance, legal, insurance, or social services.

Service Security

Do not compromise the security of Google or third parties. It is prohibited to generate or distribute content that:

• Spreads spam, phishing, or malware.
• Causes abuse, harm, or disruptions to Google’s or third-party services or infrastructure.
• Bypasses abuse protections or safety filters, including intentional model manipulation to violate these rules.

Sexual, Violent, or Hateful Content

Do not use AI to create or disseminate content that:

• Promotes hate or incites hatred.
• Encourages harassment, bullying, intimidation, or abuse.
• Promotes or incites violence.
• Contains sexually explicit elements, such as pornography or content intended for sexual gratification.

Misinformation and Deceptive Representation

It is prohibited to use AI models for:

• Fraudulent acts, scams, or other deceptive activities.
• Impersonating others (living or deceased) without clearly disclosing it, with intent to deceive.
• Promoting false expertise in critical areas such as health, finance, or legal domains.
• Spreading misinformation about democratic processes, healthcare practices, or similar topics with deceptive intent.
• Falsely claiming that AI-generated content was created solely by a human.

Exceptions

Exceptions to these limitations may be considered when there is an educational, scientific, documentary, or artistic purpose, or when the public benefits clearly outweigh the potential risks.