Artificial intelligence notice

How AI is involved, what it receives and what it cannot establish.

Version: 23 October 2026. Lira uses artificial intelligence to discuss your home and prepare reports you request. This notice explains the information used, how responses are generated and the limitations to consider.

1. When AI is involved

LiraProtect uses an OpenAI generative model to discuss your home and prepare guidance and reports based on your questions, selected photos, description and preferences. Reference documentation is optional. Answers may use relevant excerpts from the published library; they must not claim documentary support that was not consulted. You do not need to upload a prior report. You are not automatically speaking to a person or receiving human review of every answer. If a request cannot be completed, the application explains this rather than presenting a simulated answer as a real analysis of your home.

2. Data sent

Answers use your messages, relevant descriptions and preferences, recent context and reduced versions of associated photographs where present. Images may be sent again with later conversation messages. Report chat uses report text and recent context without resending photographs. With text-only input, AI must not claim to have seen images you did not supply. Generating or updating a report uses the saved conversation, including older messages, to summarise the discussion; this does not request a second photograph analysis. Reports remain independent and do not change automatically as the chat continues. Avoid faces, documents, exact addresses and unnecessary sensitive information. Image processing removes embedded metadata, not personal details visible in a scene. Read Privacy for providers, retention and rights. Cookie choices are separate from requesting an AI response or sending a photograph.

3. Sources and types of claims

Answers should distinguish visible observations, inferences, technical recommendations, perceptual effects and cultural or symbolic references. Reports may use relevant excerpts from the published library. General chat also retrieves relevant excerpts from documents published for AI use. When web search is available, Lira can check or expand public information and show provider-cited links. An inventory or bibliographic entry does not mean the full book is available or has been read. Not every response involves a search. Instructions exclude personal data and identifying home details from queries, but are not an infallible filter. Web results may be outdated or incorrect. A citation is not a guarantee of human verification, currency or relevance. Cultural symbolism is not a proven physical property.

4. Errors and limitations

AI may omit information, misunderstand photos, make errors or reproduce biases. It cannot verify hidden materials, measurements, regulations, installations or hazards. It does not certify safety or replace technical, medical, legal or emergency professionals. Health, crime prevention, protection, savings and wellbeing outcomes are not guaranteed.

5. Before applying advice

Check your needs, mobility, household, tenancy and budget. Verify measurements and stability. Seek qualified review for building work, structures, electricity, gas, fire safety, regulated accessibility and other significant risks. Stop if danger appears. Contact local emergency services in an emergency.

6. Human control and correction

You decide what to save or implement. Add context, request clarification or delete content. Professional review and a specialist network are not automatic. Report problematic answers to maryalejandrasanchezg@gmail.com without adding sensitive material to a public channel.

7. Images and examples

Ambient images are illustrative and may be AI-generated. They are not proof of completed work, client testimonials or real before-and-after results. Example assessments are labelled and do not evaluate the visitor’s files.

8. Privacy and transparency

LiraProtect does not use your photographs to train its own models. OpenAI requests include store:false, which does not itself guarantee zero retention or anonymity: the provider may retain data under applicable terms and for safety purposes. Privacy explains providers, purposes, retention and how to exercise your rights. This AI information does not remove users’ rights or legally required responsibilities.