system of artificial intelligence
Computational system making predictions, recommendations or decisions.
How the laws define this
Regulators consistently establish that an artificial intelligence system consists of computational models or software designed to process inputs and generate outputs—such as predictions, recommendations, or decisions—that influence real, physical, or virtual environments. Every instrument includes the underlying software, models, and connected components within its regulatory scope. The primary divergence lies in how these legal frameworks restrict or expand their statutory coverage. Proyecto de Ley (Boletín 15869-19) applies a strict use- and effect-based threshold, targeting only systems capable of producing legal, economic, or bodily harms to explicitly exclude generic software. Conversely, Legge n.132 del 23 settembre 2025 - Disposizioni e deleghe al Governo in materia di intelligenza artificiale adopts a broad scope aligned with EU AI Act terminology, explicitly encompassing data pipelines and datasets. Additionally, while Projeto de Lei nº 21/2020 requires systems to act for objectives defined by humans, the Proyecto de Ley que regula los Sistemas de Inteligencia Artificial centers its scope on machine-based systems inferring outputs from data.
Synthesised from the 5 statutory definitions below. Each one is quoted in full, with its source.
How jurisdictions differ
What is distinctive in each definition, binding law in force first.
| Jurisdiction | Instrument | Status | What is distinctive |
|---|---|---|---|
| Italy | Law No. 132 of September 23, 2025 - Provisions and Delegations to the Government on Artificial Intelligence | In Force | Adds explicit coverage of datasets and deployment contexts for national oversight, aligning with EU AI Act while broadening beyond core software and models. |
| Brazil | Bill No. 2,338/2023 (Legal Framework for Artificial Intelligence — Draft national AI regulatory framework) | Under Review | Emphasizes statutory scope covering software, models, and assemblies of components whose outputs influence individuals, contexts, or automated processes. |
| Chile | Bill to regulate AI systems, robotics and related technologies | Under Review | Adopts a use- and effect-based approach targeting systems capable of producing legal, economic, or bodily harms, excluding generic software below that threshold. |
| Brazil | Bill No. 21/2020 (Foundations, principles and guidelines for development and application of AI in Brazil) | Withdrawn | Specifies that the system operates for objectives defined by humans, limiting scope to purposeful deployments affecting real or virtual environments. |
| Chile | Government bill to regulate AI systems | Under Review | Highlights use of data and computational models to infer outputs like content, establishing it as the foundational scope term for all legal obligations. |
What it triggers
- In Italy, falling under this term imposes governance rules, sector obligations, and criminal sanctions for developers, deployers, and public bodies.source
- Under Bill No. 2,338/2023, it would require developers and deployers to follow risk categories, governance rules, and rights for affected individuals.source
- Under Bill No. 21/2020, it would have required adherence to foundational principles, rights, duties, and governance instruments.source
- Under the Bill to regulate AI systems, it would impose a risk-based compliance framework for developers, providers, and operators.source
- Under the Government bill to regulate AI systems, it would classify systems by risk and establish obligations for development, deployment, and use.source
Definitions (5)
A computational system that, for objectives defined by humans, can make predictions or recommendations or take decisions affecting real or virtual environments; this encompasses the software, models and connected components considered by the bill for regulatory duties and obligations.
A comprehensive concept covering models, algorithms, software and the associated data pipelines and components that together perform tasks or produce outputs by means of automated decision‑making or content generation; the definition aligns with and adopts the EU AI Act terminology while encompassing datasets and deployment contexts relevant for national oversight.
Computer systems designed to produce predictions, recommendations or decisions that can affect the virtual or real environment; the statutory scope covers software, models and assemblies of components whose outputs can influence individuals, contexts or automated processes.
A functionally defined set of software systems that, given input data, produce outputs (e.g., predictions, recommendations, or decisions) capable of influencing physical or digital environments. The definition is use- and effect-based to target systems whose operation can produce legal, economic or bodily harms and to exclude generic software not meeting that threshold.
A machine-based system that, using data and computational models, infers outputs such as predictions, content, recommendations or decisions from given inputs; the definition covers systems whose outputs can influence physical or virtual environments and is the foundational scope term for obligations under the law.
Not to be confused with
This term explicitly covers datasets, deployment contexts, and national oversight elements, while the related term focuses solely on machine-based inference from inputs to outputs.
This definition widens to include data pipelines, datasets, and deployment contexts beyond the related term's basic input-output production per EU AI Act.
Questions and answers
What does Italy's Law No. 132 include in its AI system definition?
It covers models, algorithms, software, data pipelines, and components for automated decision-making or content generation. The definition adopts EU AI Act terminology and adds datasets and deployment contexts for national oversight.
How does Chile's Bill to regulate AI systems narrow its AI definition?
It defines AI as software systems producing outputs that influence environments, but uses a use- and effect-based approach to target only those capable of legal, economic, or bodily harms, excluding generic software.
What obligations arise from falling under these AI definitions?
In Italy, it triggers governance rules and criminal sanctions. In Brazil and Chile bills under review, it would impose risk-based duties, principles, and compliance frameworks depending on the instrument.
Related Terms
Intelligent system (AI system)
Automated or semi-automated systems affecting human decisions or outcomes....
Sistemas de inteligencia artificial
AI systems covered by the draft law....
Sistema de inteligencia artificial
Machine-based system that produces outputs from inputs, per EU AI Act....
Systems of artificial intelligence
Automated systems that perform predictions, recommendations or decisions....
Artificial Intelligence System (AI System)
A machine-based system inferring from input to generate outputs like predictions or decisions influencing environments....