Gaeilge Español

AI & Automation Disclosure

Ensuring that people are made aware, at the point of use, of the presence of any digital systems acting autonomously – that any such systems operate in the public interest – and that human oversight is used when making decisions that affect people

AI & Automation Disclosure means ensuring that people are made aware when digital systems are acting on them, shaping what they see, influencing decisions, or communicating in ways that appear human.

We believe that public trust depends on people being able to tell when they are dealing with a machine, what that machine is doing, and who remains accountable for its actions and outcomes.

For a public-serving entity, AI & Automation Disclosure means clearly disclosing where and when automated systems are used, especially when they influence recommendations, moderation, eligibility, prioritisation, or access to services. It means ensuring that significant decisions affecting people remain subject to meaningful human oversight and review, and that automated systems can be challenged, explained, and corrected where necessary.

AI & Automation Disclosure in the public interest does not resist innovation. But it does ensure that when systems act on people’s behalf or in place of people, those systems remain transparent to – and contestable by – the public they serve.

Example requirements (illustrative)

These example requirements are grounded in established international standards, regulations, and laws, which are listed in full in the section below.

  • AI systems, automated decision-making, and synthetic media are clearly disclosed to people at the point of interaction or use.

  • Synthetic or AI-generated text, audio, video, or images presented as factual or authoritative material is clearly identified as such.

  • High-risk AI systems undergo formal risk, privacy, and safety assessments before deployment and at regular intervals thereafter, with clear explanations of purpose, capabilities and limitations published for public-facing services.

  • Organisations maintain documented governance processes for the procurement, deployment, testing, and review of AI systems, designed to reduce bias, discrimination, unsafe outcomes, and foreseeable harm.

  • Automated systems affecting people operate under meaningful human oversight, with clear routes for review or intervention.

  • People can challenge or appeal significant decisions made or assisted by automated systems.

Standards, regulations, and laws informing this work

Organisations working in this area

Related standards: Responsibility to the Future Interoperability