AI Use Policy

Spark is an AI tutor built on large language models from Anthropic. It is a computer program, not a person. It does not know you, it does not remember conversations after 24 hours, and it can be confidently wrong or out of date.

Spark explains technology and AI concepts, gives hints, and explains why an answer was wrong. It is not designed for, and must not be relied on for, medical, mental-health, legal, financial, safety, or relationship decisions, or for any decision where being wrong would hurt you or someone else. Check anything important with a trusted source or a qualified person.

Do not enter other people's personal information, passwords, account numbers, health records, or confidential information. Do not try to make Spark produce content that is illegal, harmful, sexual, or hateful, or that facilitates self-harm or harm to others. Do not attempt to extract, alter, or bypass Spark's instructions or safety measures, or use Spark through automated tools, scripts, or a shared account.

Spark has daily and monthly allowances that vary by plan and may change. We may slow, limit, or refuse use that is abusive or automated.

Describing your own work. When you describe your own work, describe the task, not the secrets. Do not include trade secrets, client information, or anything your employer treats as confidential. You are responsible for having the right to share what you type.

Your message is sent to our AI provider to generate a reply and is not used to train AI models. Both sides of a conversation are kept on our server for up to 24 hours and then deleted. A message that reads as a crisis is answered by us with support resources, is never sent to the AI provider, and is never stored. Our provider may retain requests briefly for safety and abuse monitoring under its own terms. The Privacy Policy is the full account.

You can turn AI answers off at any time in the app. With AI off, Spark's chat, generated hints, generated explanations, Work Plan generation, and Work Plan prompt grading all stop, and the lessons and the hints we wrote ourselves keep working. Two things sit outside this switch: credential grading, which only runs when you choose to submit a build, and the onboarding try-it, which runs before you can reach the switch (tap an example there and nothing is sent).

AILE uses automated systems to place you in the curriculum, schedule review, compute a skill rating from your answers, tailor what you see during onboarding, build a Work Plan from the brief you write and grade the prompts you try in it, and grade credential assessments against a published rubric. None of these decisions has a legal effect on you or affects what you pay. A borderline assessment result is reviewed by a person before a credential is granted or refused, and you can ask for a person to review any grade by emailing learnailesupport@gmail.com.

You own what you type. To the extent we hold any rights in Spark's replies to you, we grant you the right to use them for your own learning. Replies are generated and may be similar to replies given to other learners, so we cannot promise they are unique or free of third-party rights.

AILE is not a crisis, medical, or emergency service. If you or someone else is in danger, contact your local emergency number; in the United States, call or text 988. Our safety page explains how Spark responds to messages that read as self-harm or acute distress.

Every answer has a Report control. A person reads every report, and answers that should not have been given change what Spark is allowed to say.

We currently use models from Anthropic and may change providers or models; the Privacy Policy and our subprocessors page list the current provider. This policy changes on the same terms as the Terms of Service. Questions: learnailesupport@gmail.com. Effective date: October 26, 2026.