Human responsibility
We use AI for clearly defined tasks. A person or designated role remains responsible for every use. Critical outputs are reviewed according to risk and are not adopted as facts or decisions without review.
- Human responsibility is not delegated to a model or vendor
- Use and oversight reflect the purpose, potential effects, and people affected
- Users are informed clearly and accessibly when they interact directly with AI
Data, confidentiality, and rights
We use AI to process only data that is necessary for the relevant purpose and permitted by law and contract. Personal, customer-related, or confidential information is processed only in systems approved for that purpose. We consider copyright, trademark, personality, and confidentiality rights for both inputs and outputs.
Quality, security, and limitations
AI systems can be wrong, outdated, or produce unexpected results. We therefore review outputs according to their purpose and risk and restrict access and permissions. Technical and organisational safeguards are selected to suit the relevant use.
Skills and keeping current
People who use or oversee AI for EA receive information appropriate to their role and context about capabilities, limitations, privacy, security, copyright, and escalation routes. We update this knowledge and our rules after material changes and at least annually. The EU AI Act does not prescribe a particular individual certificate or uniform proficiency level for this purpose.
Case-specific review
These principles apply to EA’s general use of AI. Each customer project is also reviewed from a legal, technical, and organisational perspective in light of its specific purpose, data, risks, and contractual requirements.