AI and Automated Processing Notice
This notice supplements the Privacy Policy and Terms. SocialGryd generates user-facing food estimates, workout plans and coach guidance using OpenAI. These are automated outputs, not professional assessment or medical/nutrition advice.
1. Inputs and outputs
Food requests can include photographs, packaging/labels, descriptions, clarifications, units, locale, barcodes and selected previously saved or corrected meal records. Product catalogue lookup and web search can provide supporting information. Workout requests can include your instructions, goals, experience, activity preferences, recent workouts, saved plan, programme progress and check-in feedback. A hashed account safety identifier accompanies API requests. Free text and images can contain identifying or health information even without a name field.
The Privacy Policy describes the selected-history limits. Some outputs are saved to your private account for subsequent use. Accepting a suggested plan does not certify medical suitability. You can ignore a suggestion or correct an estimate.
2. Fitness and nutrition safety
AI can invent information, misunderstand images and suggest unsuitable exercise or nutrition values. Calories, macros, portions, energy expenditure and targets are estimates. Do not use them for allergy safety, medication dosing, clinical diets, diagnosis, treatment or unsafe restriction of food. Do not skip necessary nourishment or overexercise to satisfy a number.
Review every workout, stay within safe abilities, follow qualified medical advice and stop for concerning symptoms. If unsure whether an exercise or dietary change is safe, do not proceed until you obtain appropriate advice. The detailed safety requirements and non-excludable liability protections in the Terms apply to every AI feature and website calculator.
AI food estimates and AI coaching are available only to adults aged 18 or over during beta, with separate data consent. No guardian-consent exception is currently offered for these features.
3. Provider handling
Current requests use OpenAI's API with response storage disabled. This does not remove all provider retention, including abuse-monitoring and applicable safety/legal exceptions. SocialGryd does not authorise general foundation-model training on private user content. Provider account choices and terms must be maintained consistently with this restriction. See the Provider Register and OpenAI data controls.
AI food and coaching obtain separate recorded informed consent before use, including explicit consent to health-data processing. Terms acceptance and device camera permission do not substitute for explicit health-processing consent. Withdraw either AI choice in Settings > Data and privacy. Data requests can also be made at legal@socialgryd.com.
4. Ranking, moderation and other automation
Other automated processing supports feed/discovery ordering, proximity matching, activity totals, challenge scores and abuse checks. Relevant inputs may include recency, follows, selected interests, interactions, location where permitted and safety signals. Google Cloud Vision supports image safety checks. An automated score or match is not identity verification or a judgement of your health or worth.
Automated content checks or restrictions may be wrong. Contact support@socialgryd.com to challenge a restriction, identify the relevant content or account and explain the error. We provide human review where required. This notice does not claim that every moderation action has already received human review or that an AI-system regulatory classification has been certified.
5. Updates
New AI uses or providers require corresponding notices and any required consent. The Privacy Policy explains changes and rights. Former references to Anthropic categorisation and speculative AWS systems are not descriptions of the current reviewed request paths.