EU AI Act transparency

Purpose of this notice

DAWN includes AI functionality. This notice explains the intended boundaries and responsibilities relevant to the EU AI Act. It is not a declaration of conformity, certification or claim that every configuration satisfies every regulatory requirement.

Transparency in interaction

Conversational output and generated speech are produced by AI and can be mistaken. People should know when they are interacting with an AI system. Recording and transcription are separate processing activities and require their own notice and lawful basis. Website illustrations and recorded demonstrations do not make decisions about visitors.

Employment use needs specific assessment

AI used to recruit or select people, evaluate candidates, allocate work or influence employment decisions can fall within the Act’s high risk provisions. Calling a tool an assistant or adding a human confirmation step does not by itself remove that classification.

The intended purpose, actual influence, data and deployment context must be assessed before the use case is put into service. The applicable legal obligations and implementation dates must be checked for that use case. We do not treat every DAWN feature as having the same risk classification.

Provider and organisation responsibilities

DAWN is responsible for obligations applicable to its role in the system it supplies. Organisations are responsible for obligations applicable to their deployment, including appropriate oversight, staff competence, input data, required notices and monitoring. Neither party can transfer away a statutory duty simply by contract.

Where required, the use case needs documented risk management, technical information, records, impact assessments and procedures for affected people. Do not use a high risk configuration merely because a demonstration is available. Prohibited uses remain prohibited regardless of a customer instruction.

Questions and concerns

Contact hello@dawnhcm.com to discuss a specific intended use, required documentation or a concern about AI output. Identify the task and the effect on people rather than sending personnel records. Affected individuals should also contact the organisation responsible for the relevant employment decision.