On the proposal for a regulation of the European Parliament and of the Council laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts
ENVI-PA-699056
Committee on the Environment, Public Health and Food Safety · Rapporteur: Susana Solís Pérez
Written by a language model from the full text only; every figure comes from the text. The paragraph links open the paragraph a statement rests on — check the text itself before relying on it.
The Committee on the Environment, Public Health and Food Safety proposes amendments to the proposed Artificial Intelligence Act. It adds the environment to the areas needing a high level of protection, so AI systems harming the environment become high risk.¶¶¶ It creates a definition of 'end recipient' and gives end recipients rights to transparency and information, alongside users.¶¶¶ It adds health and mental health protections, including a new recital on the WHO definition of health and the right to physical and mental health.¶¶¶ It requires AI systems to respect the 'do no significant harm' principle and the Aarhus Convention redress rights for environmental harm.¶¶¶ It calls for alignment with other laws such as the GDPR and the European Health Data Space, and for the Board to cooperate with sectoral advisory groups.¶¶¶
Committee position. The rapporteur proposes amendments to the AI Act to add the environment as a protected area, create a definition of end recipients with information rights, strengthen health and mental health protections, and align the regulation with other EU laws.¶¶¶
Key points
- The environment is added alongside health, safety and fundamental rights as an area requiring a high level of protection throughout the regulation.¶¶¶
- AI systems that can have major negative implications on the environment are classified as high-risk AI.¶¶
- The 'do no significant harm' principle from the Taxonomy Regulation is set as a limit for AI systems.¶
- The right to redress for negative environmental impacts under the Aarhus Convention is reinforced.¶
- A new definition of 'end recipient' is added, covering any person other than an operator to whom an AI system's output is intended or provided.¶¶
- End recipients are granted transparency and specific information rights, including on system outputs, accuracy and residual risks.¶¶¶
- New recitals state that health includes mental health, that the right to physical and mental health is a fundamental human right, and that AI can harm or help mental health.¶¶
- The regulation should promote health, wellbeing, disease prevention and healthy lifestyles.¶
- AI systems intended to distort human behaviour and cause psychological harm or disruption of the sense of oneself are forbidden.¶
- High-risk classification criteria are extended to cover risks to the environment and biodiversity, including irreversible environmental outcomes.¶¶¶
- Providers must report serious incidents or malfunctions that breach Union law protecting health, safety, fundamental rights and the environment.¶
- The Board and national authorities must cooperate with civil society, advisory and expert groups to ensure a consistent approach.¶¶¶
Who is affected
- Providers of high-risk AI systems, who must meet transparency, reporting and risk management duties.¶¶
- End recipients such as patients, students and consumers, who gain information and transparency rights.¶¶¶
- National competent authorities, which must have expertise in environmental risks and adequate resources.¶¶
- The healthcare sector, where AI is to be used with interoperable health data and sandboxes.¶¶¶