To use our API, every developer must clearly outline their use case and have it approved by Cohere through our application process. The application requires thoroughly understanding of our models and their limitations, which will be updated as the models improve. Beyond these Usage Guidelines, you should refer to the Generation and Representation model cards for detailed information about each model.
By understanding the language models that power our API endpoints, being aware of their limitations, and documenting your development practices, you can do great things with the Cohere API.
Using the Cohere API may accelerate application development considerably, but requires careful consideration of the potential harms of your system, as well as documentation of relevant data, models, risks, and biases. Here are some steps you might take toward these ends:
- Make model and dataset information accessible to your users, investigating and clearly documenting any biases along demographic and intersectional dimensions
- Provide a link to the corresponding Cohere model card, explaining how your application uses its output
We do not permit the usage of our API for any of the following purposes. These examples are illustrative but not exhaustive; Cohere reserves the right to terminate access for harms which are not listed at our sole discretion.
- Violence and threats:
- Violence/Incitement: Actions that threaten, encourage, or incite violence against anyone, directly or indirectly.
- Self-harm: Promoting or glorifying acts of self-harm, such as cutting, eating disorders like anorexia or bulimia, and suicide.
- Sexual exploitation: Promoting or celebrating sexual exploitation, including the sexualization of minors.
- Hate speech: Promoting hatred or glorifying abuse against people based on characteristics like race, ethnicity, national origin, religion, disability, disease, age, sexual orientation, gender, or gender identity.
- Antisocial and antidemocratic uses:
- Harassment: Bullying, threatening, shaming, doxing.
- Insensitivity: Belittling victims of serious physical or emotional harm (even if unintentional).
- Intentional sowing of division: Sharing of divisive generated content in order to turn a community against itself.
- Harmful belief perpetuation: Perpetuating racism, sexism (even if unintentional).
- Applications that aim to characterize identity: Attempting to characterize gender, race, ethnicity.
- Graphic depictions: Distribution of sexually explicit acts, torture, or abuse.
- Fraud: Catfishing, phishing.
- Spam: Sending unsolicited email and messages or manipulating search engines.
- Misrepresentation: Representing raw generations as coming from humans, using supervised generations with false identities, or a single person using generations with many identities that appear to be independent.
- Misinformation: Creating or promoting harmful false claims about government policies or public figures.
- Intentional manipulative redirection of attention: Sharing positive generated content in order to direct attention away from harmful actions.
- Attacks on security or privacy:
- Security breaches: Spearphishing.
- Privacy violations: Model attacks to extract personal information.
- Unsafe unsupervised uses: Usage in an unsupervised way for medical or mental health purposes, or other sensitive use cases which may cause harm to humans.
Usages which appear to violate our guidelines should be reported within 24 hours to Cohere by contacting us at email@example.com.
Note about adversarial attacks: Intentional stress testing of the API and adversarial attacks are allowable, but violative generations must be reported immediately, and must not be used for any purpose except for documenting the result of such attacks in a responsible manner.