AI & Ethical Use Policy

Last Updated: May 26, 2026

GlucoSpike AI uses artificial intelligence to help you make more informed food choices. This page explains exactly how AI is used in the app, what it can and cannot do, and how we approach AI responsibly and ethically.

1. How AI Is Used in GlucoSpike AI

Meal Analysis

When you photograph a meal, the image and your behavioral profile are sent to Google Gemini, a large multimodal AI developed by Google. Gemini returns:

A GlucoScore (1–10) — an estimate of how likely the meal is to cause a blood glucose spike. A higher score means a better choice for blood sugar management.
Nutritional estimates — approximate calories, carbohydrates, protein, fat, and fiber (these are AI estimates for unpackaged food; packaged products use verified database values — see below)
Identified ingredients
Meal insights — observations about what makes the meal a good or less ideal choice
Suggested swap ideas and questions you might have about the meal
A post-meal walk duration suggestion (in minutes)

Product Barcode Scan

When you scan a packaged product barcode, GlucoSpike AI looks up the product’s actual nutritional data from Open Food Facts — a public, nonprofit food database with verified data from manufacturers and contributors worldwide. This means the calorie count, carbohydrates, fiber, and ingredients you see for packaged foods come from real product data, not AI estimates.

That verified product information is then analyzed by Google Gemini to generate:

A GlucoScore
A verdict (good pick / use with caution / avoid)
Key concerns and a tip

Barcode scan results are not saved to your account — each scan is processed in real time and discarded.

2. What the AI Sees

The only personal context sent to Gemini is your onboarding quiz answers — general lifestyle information such as your typical energy patterns, meal timing, activity level, and daily carb target. This is the full extent of what the AI knows about you.

GlucoSpike AI is intentionally designed not to collect medical or clinical data. We do not collect diagnoses, medications, blood glucose readings, biometrics, or any clinical health metrics — so none of that data exists in the app to begin with, let alone reaches the AI.

3. AI Limitations — What GlucoScores Are Not

AI-generated GlucoScores are wellness estimates, not clinical predictions. They should not be treated as medical measurements.

Important limitations:

Portion size uncertainty: The AI estimates portions from photos. Actual portion sizes may differ significantly.
Preparation variation: Two “grilled chicken salads” may have very different glycemic impacts depending on dressings, cooking methods, and ingredients.
Individual variation: Blood glucose responses to food vary significantly between individuals based on gut microbiome, sleep, stress, and metabolic health. A GlucoScore of 8 does not guarantee a stable glucose response for every person.
Photo quality: Poor lighting, unusual angles, or mixed dishes can affect recognition accuracy.
AI model limitations: Like all AI systems, Gemini can make errors, particularly with unfamiliar foods or cuisines.

GlucoSpike AI is not a substitute for:

A continuous glucose monitor (CGM)
A registered dietitian or nutritionist
Your doctor or healthcare provider
Any prescribed medical treatment or dietary plan

4. What Data Is Sent to the AI

Feature	Nutrition Source	Data Sent to Google Gemini
Meal photo analysis	AI-estimated (approximate)	Compressed meal photo + behavioral quiz answers
Product barcode scan	Open Food Facts database (verified product data)	Product name, brand, nutrition per 100g, ingredients text — no personal data

What is never sent to the AI:

Your name, email, or any directly identifying information
Any other user’s data

5. AI Data Retention (Google’s Policy)

As of May 2026, Google’s Gemini API does not use API inputs to train its foundation models by default under its enterprise API terms. This means meal photos and behavioral context you send for analysis are not used by Google to train Gemini.

This is subject to Google’s policies, which may change. We encourage you to review Google’s Gemini API terms for the most current information. Our Privacy Policy also covers this.

6. Transparency & Disclosure (EU AI Act)

GlucoSpike AI is designed to be transparent about when you are interacting with an AI system.

AI interaction disclosure: Every GlucoScore and product scan result you receive is generated by an AI system (Google Gemini). This is disclosed within the app and in these policies.
Risk classification: Under the EU AI Act, GlucoSpike AI is assessed as a minimal to limited risk AI system. It makes wellness estimates about food choices — not clinical diagnoses, and not consequential decisions about your medical care.
No high-risk AI use cases: We do not use AI for medical condition profiling, biometric identification, or any purpose that would classify as high-risk under the EU AI Act.
No automated decision-making with legal effect: GlucoScores are informational estimates. No automated decision-making with legal or significant personal consequences is performed.

7. Our Ethical Commitments

Accuracy over engagement: We display GlucoScores and insights as estimates with appropriate context, not as definitive health verdicts. We do not gamify AI scores in ways that could encourage unhealthy behaviors.

No medical overreach: We have deliberately excluded all medical condition collection from the app. The AI provides food context — it does not diagnose, predict disease, or act as a clinical tool.

Data minimization: We send only what is needed to generate useful results. We do not send your name, email, or any identifying information to the AI. Barcode scan data is not stored.

User control: You can delete your account and all associated data at any time. Notification preferences are fully under your control.

Third-party accountability: We use Google Gemini under enterprise API terms that include data protection commitments. We monitor changes to Google’s AI policies and will update our practices accordingly.

8. Feedback & Concerns

If you believe a GlucoScore is inaccurate, or if you have concerns about how AI is used in the app, please contact us:

Email: support@glucospike.ai

We take AI accuracy and ethical use seriously and welcome feedback from our users.