Acceptable use policy

What the AI features are for

AI writing feedback scores the answers you write to practice writing tasks and explains how to improve them.

AI speaking examiner plays the examiner in practice speaking tasks, replies to what you say, and scores the transcript of your answers.

Use them to practise genuine exam-style answers in French: introducing yourself, asking for information, giving and defending an opinion, writing messages, articles and letters. Anything else is outside what these features are built for.

Sexually explicit and NSFW content is prohibited

You must not submit, request or try to generate sexually explicit, pornographic or otherwise NSFW (not safe for work) content of any kind in the AI writing or speaking features. This includes erotic stories or role-play, graphic sexual descriptions and sexual content presented as a “practice answer”.

Any content that sexualises minors is absolutely prohibited. We have zero tolerance for it: the account is closed immediately and we report it to the relevant authorities.

Other prohibited uses

You must not use the AI features to submit or generate:

Hate or harassment: content that attacks, demeans or threatens people because of race, ethnicity, national origin, religion, sex, gender identity, sexual orientation, disability or any similar characteristic, or that bullies or harasses anyone.

Violence and self-harm: threats of violence, content that glorifies or promotes violence or terrorism, or content that encourages suicide, self-harm or eating disorders. If you are struggling, please contact local emergency services or a crisis line; in Canada you can call or text 9-8-8.

Illegal activity: content that promotes, plans or facilitates illegal acts, including fraud, immigration fraud and document forgery.

Other people’s personal data: names, contact details, ID or passport numbers, immigration file numbers or other personal information about someone else, and sensitive information (health, financial or government ID details) that the task doesn’t need. Invent details for practice answers instead.

Malware, spam and abuse: malicious code, phishing text, spam or content meant to deceive people.

Content you don’t have the right to use: other people’s copyrighted work submitted as your own, or material you are not allowed to share.

Misusing the service itself

Manipulating the AI: don’t try to jailbreak the AI, override its instructions, make it ignore its role, extract its prompts or system instructions, or make it produce anything other than exam practice. Your answers are treated as text to assess, never as instructions.

Automated or bulk use: don’t use bots, scripts or scrapers, or send automated or high-volume requests to the AI features or any other part of the site.

Reselling or sharing: don’t resell, sublicense or give others access to the AI features, and don’t share your account or Premium access.

Cheating and impersonation: don’t use the features to get help during a real exam, and don’t present NCLC French, its feedback or its scores as official results or as coming from France Éducation international, Le français des affaires (CCI Paris Île-de-France) or IRCC.

Our safeguards

Instructions to the AI: the AI is told to stay in its exam role, to treat your text as data to assess rather than instructions, and never to produce sexual, explicit or other prohibited content. If a submission contains such content, it is not evaluated: you get a score of zero and a short note that the content breaks this policy. The speaking examiner declines to engage and steers the conversation back to the exam task.

Our AI provider’s safety systems: the AI features run on models from Anthropic, which applies its own usage policies and automated safety filtering. Requests that break those policies can be refused; when that happens you get a basic, clearly labelled estimate or the scripted examiner instead of AI feedback.

Limits on input: answers are capped in length (about 12,000 characters per writing task and 1,500 characters per spoken turn), a request covers at most three tasks, and the speaking examiner only runs the role-plays from our own practice tests.

Rate limits and allowances: each connection can make a limited number of AI requests in a short period (for example 10 evaluations every 10 minutes), free use is capped per account or device and per network per day, and the extended AI features require a Premium account.

Security logs: we keep standard server logs, as described in our privacy policy, so we can investigate abuse.

What happens if you break this policy

Depending on how serious the violation is, we may refuse to evaluate the content, limit or pause your access to the AI features, or suspend or close your account.

For serious or repeated violations we may close the account without a refund, regardless of the refund guarantee.

We report content that sexually exploits children, credible threats of violence and other illegal content to the relevant authorities where the law requires or allows it.

Report misuse

If you see the AI produce something inappropriate, or you think someone is misusing NCLC French, email [email protected] with a short description and, if you can, the date and the page. We read every report.

We may update this policy; the date above shows the latest version. Please also read our terms of use.