AI Policy

Veld Wyser uses artificial intelligence, data models, and environmental data to support decision-making in livestock and land management.

This policy explains how these systems work, their limitations, and the responsibilities of users.

Our AI systems are designed to:

• estimate vegetation and biomass;
• calculate grazing capacity (e.g. LSU/ha);
• analyze environmental conditions;
• read ear tags and identify plants from photos;
• transcribe voice messages and summarize minutes in study groups;
• provide management recommendations.

These systems are intended to support, not replace, human decision-making.

AI outputs are based on a combination of:

• satellite data (e.g. NDVI, NDRE);
• weather data (rainfall, temperature);
• user-provided data;
• ecological models.

The quality of outputs depends on the quality, resolution, and availability of these inputs.

Users acknowledge that:

• AI outputs are probabilistic and not deterministic;
• environmental systems are highly variable;
• data may be incomplete, delayed, or inaccurate;
• models rely on assumptions that may not apply in all situations.

Rangelands are complex systems influenced by rainfall, soil, grazing pressure, and management practices. As a result, outputs may vary significantly across time and location.

The AI outputs provided by Veld Wyser do not constitute:

• agricultural advice;
• financial advice;
• operational guarantees.

Users must independently validate all decisions.

Users remain fully responsible for:

• livestock management decisions;
• grazing strategies;
• financial outcomes;
• compliance with local regulations.

We continuously improve our models using aggregated and anonymized data. This may lead to changes in outputs over time.

We aim to provide clear explanations of how outputs are generated, including:

• key input variables;
• model assumptions;
• confidence indicators (where available).

8.1 You always know when you are dealing with AI. In line with article 50 of the EU AI Act, AI-generated content is labelled as such. The AI advisor is presented as an assistant and never as a human adviser.

8.2 Classification under the EU AI Act. Veld Wyser is a decision-support tool for grazing and land management. We do not consider it a high-risk AI system within the meaning of Annex III: it is not used for access to essential services, employment, creditworthiness, or law enforcement. It falls under the transparency obligations for AI systems that interact with people.

Should we ever apply AI in a context that does qualify as high-risk — for instance linked to financing or insurance — we will reassess that classification and inform you beforehand.

8.3 Models used. The AI features in the app run on models from OpenAI: the AI advisor, the transcription of voice messages and the summaries of minutes in study groups, and the photo recognition of ear tags and transect photos. Vegetation and biomass analyses use our own models based on Sentinel-2 satellite data. OpenAI acts as our processor and does not use this data to train its models. See section 6 of our Privacy Policy for which data is sent to OpenAI.

By using the Service, users acknowledge that:

• reliance on AI outputs involves risk;
• no model can fully capture real-world ecological complexity;
• decisions should not be based solely on automated outputs.

For questions about our AI systems: info@veldwyser.com