At Montala we believe in the responsible and ethical use of Artificial Intelligence (AI) to enhance both our internal productivity and the services we provide. This policy outlines our principles and practices relating to AI, covering internal use, customer impact, environmental sustainability, and ethical considerations.
Internal Use of AI
We encourage our team to use AI tools to improve productivity and creativity, while ensuring that all work produced meets the high standards our customers expect.
- Code and Technical Work: When using AI to assist with coding, team members must fully understand, test, and take responsibility for any code that is added to the ResourceSpace base — exactly as if it had been written independently.
- Content and Communications: AI can support content creation by generating ideas but team members are responsible for verifying accuracy and appropriateness and ensuring all blogs, documentation and marketing materials align with our tone, values and standards.
- Decision Support: AI can aid decision-making but must never replace expert human judgment. Information provided by AI must be independently verified using reliable sources, especially since AI can sometimes generate inaccurate or fabricated content ("hallucinations").
- Accountability and Standards: AI is a tool — not a substitute for professional knowledge or integrity. All final work must reflect the standards and values of Montala.
AI and ResourceSpace Functionality
We integrate AI features in ResourceSpace to help users save time and enhance asset management, including tools for automatic tagging, image similarity, and summarisation.
- User Control and Transparency: AI-powered features are always optional and clearly labelled. Users are informed when AI is involved in processing data, and we aim to make its role transparent.
- Privacy and Data Security: We do not allow user-uploaded content (including images and metadata) to be used to train AI models, whether our own or third party. When AI tools are used, whether hosted by us (e.g. CLIP, InsightFace) or third parties like OpenAI GPT and Google Vision, we ensure that content is processed only for the requested task and not retained or reused for training.
See OpenAI’s data usage policy and Google’s Cloud Vision privacy terms. - Models We Host Ourselves: Some AI features in ResourceSpace are powered by models we run entirely in-house. These models (such as CLIP for similarity and InsightFace for face matching) are not connected to the internet and do not send data externally.
Environmental Responsibility
We recognise the environmental concerns surrounding AI, particularly the energy demands of training and running large models. At Montala, we are committed to reducing the environmental impact of our technology.
- Sustainable Hosting with Microsoft Azure: OpenAI GPT runs on Microsoft Azure’s cloud platform, which has been carbon neutral since 2012.
- Sustainable Infrastructure: Where we run AI models on our own servers, these are powered by 100% renewable energy, reflecting our commitment to environmental responsibility.
- Efficient AI Use: We avoid training large AI models ourselves — which is typically the most energy-intensive aspect of AI — and focus instead on integrating pre-trained models in efficient and targeted ways.
- Ongoing Commitment: As the environmental impact of AI becomes clearer, we will continue to monitor developments and prioritise low-impact, sustainable approaches wherever possible.
Ethical Considerations
AI must be used in a way that respects user rights, protects privacy, and avoids harm.
- No Use of Customer Data for AI Training: We do not, and never will, use customer-uploaded data (images, text, metadata) to train AI models.
- Fairness and Bias Mitigation: Where we use models that involve facial recognition or language analysis, we are aware of the risks of bias. We test and monitor our AI tools to minimise unfair outcomes and provide opt-out mechanisms for features such as face matching.
- Transparency and User Consent: We are committed to clear communication about how AI is used within ResourceSpace. We believe users should be informed and in control, and we welcome questions or concerns about our AI integrations.