Updated 06.05.2026
Runbox recognizes Artificial Intelligence (AI) systems and Large Language Models (LLMs) as valuable tools with the potential to improve our knowledge and capabilities in serving our customers. We also acknowledge that AI/LLMs present significant challenges and risks, and can potentially be intrusive. Therefore we commit to using AI/LLMs responsibly, ethically, and transparently in alignment with our core values and customer expectations.
Purpose
AI/LLMs will only be used as tools to augment human capabilities, never to replace human judgement. They may assist with research (analyzing public information or documentation), content creation (drafting articles or reports), and illustrations (graphics or other visual aids) provided all content is reviewed and validated by a human before publication or distribution. Our chatbot, Boxy, is powered by an AI/LLM system to provide automated responses to common support questions. However, AI/LLMs are not used to replace direct customer support communication through email or our Support Center. See Support below.
Data Privacy and Confidentiality
We are committed to protecting personal and sensitive data. All use of AI/LLMs complies with data protection regulations, including the Norwegian Personal Data Act and GDPR. No personally identifiable information, proprietary data, or confidential customer information is ever input into AI/LLMs. All AI/LLM usage is reviewed by authorized staff to ensure compliance with privacy and security standards.
Ethical Use
All AI/LLM-generated content undergoes human review to ensure accuracy, appropriateness, and alignment with our ethical standards. AI/LLMs will never be used to generate or propagate misleading information, infringe on intellectual property rights, or promote harmful, discriminatory or offensive content. AI/LLMs and generated content will be scrutinized by humans to ensure it does not reinforce harmful biases or provide inaccurate, discriminatory, or offensive outputs.
Human Oversight
AI/LLM tools are only used as assistive tools. Content generated by AI/LLMs will be reviewed and validated by authorized staff to ensure that it aligns with our standards of accuracy, quality, and ethical guidelines. This review process helps mitigate biases, errors, or misinterpretations, ensuring the final output meets our ethical and quality standards.
Support
An essential part of Runbox’s services is to provide personal, human support. Any direct communication you initiate with our customer support team (e.g., via email or the Runbox Support website) will always be handled by a human, ensuring personalized and direct interaction.
To complement this, we have created Boxy, our AI-powered chatbot, which operates separately from our human support team. Boxy is available 24/7 to provide quick answers to common questions about Runbox’s services, setup, and troubleshooting. Boxy does not replace human support — it is a standalone tool designed to assist with well-documented issues. For complex, sensitive, or account-specific inquiries, our human support team remains your best resource.
Boxy draws information only from our own knowledge base and support resources. Boxy has no memory and does not learn from conversations. It has no access to Runbox accounts, emails, or personal data. It is not browsing the web or cross-referencing information between users. It cannot remember anything you told it in a previous session. Boxy’s purpose is simply to provide quick responses to questions so customers can get help fast, day or night.
Boxy’s responses are logged for 30 days (stored securely in Norway) and reviewed regularly to ensure accuracy and alignment with our standards, but your input is never stored. Read the entire Boxy policy document here.
Read more: AI, LLMs and Boxy
Intellectual Property
Any content generated by AI/LLMs is owned by Runbox, as the entity that initiated the creation of the work. AI/LLMs are not recognized as legal authors and cannot hold legal authorship, and therefore the copyright for generated content defaults to Runbox, in accordance with applicable copyright law. We will ensure that AI/LLM-generated content does not infringe on third-party intellectual property rights.