Explainability

The ability to provide an intelligible account of how algorithmic outputs relate to decisions in a specific case, enabling individuals to understand the basis and determinative factors behind a decision.

The capacity of algorithmic systems to provide interpretable, comprehensible information about how decisions or outputs are produced, enabling transparency, oversight and user understanding as emphasized by the strategy's ethical and accountability objectives.

A compliance requirement that AI systems be capable of providing understandable explanations of their decision‑making processes to affected individuals and regulators, enabling transparency, accountability, and meaningful review of automated decisions.

A trustworthiness attribute describing the extent to which an AI system's internal mechanics, decisions, or outputs can be made understandable to relevant stakeholders through interpretable models, explanations, or representations. Explainability enables scrutiny, helps identify sources of error or bias, and supports accountability and informed human oversight.

The extent to which an AI system’s outputs and the factors influencing them can be understood and inspected by human experts, enabling practitioners to assess reliability, justify decisions to clients, and meet professional and regulatory transparency obligations.

The capacity to explain an AI system's logic, the significance of its processing, and the anticipated consequences for data subjects; the advisory requires explainability measures sufficient to inform affected individuals and enable contestability of automated decisions.

The capability of an AI system to provide understandable, relevant reasons for its outputs to affected stakeholders, supporting transparency, meaningful challenge, and accountability; encompasses both user-facing explanations and internal interpretability/documentation proportional to impact.

The extent to which an organisation can provide understandable reasons for AI-driven outputs, including the appropriate level and form of explanation tailored to different stakeholders; explanations should be proportionate to assessed risk and enable users to understand, contest or seek redress for decisions.

Information and mechanisms designed to make AI-driven decisions and processes understandable to affected individuals and stakeholders, encompassing both how a system makes decisions (process explanations) and why particular outcomes were reached (outcome explanations), to support transparency, contestability and data subject rights.

The capacity of an AI system to provide meaningful, interpretable information about its internal logic, decision-making processes and outputs, sufficient to allow users and regulators to understand capabilities, limitations and the provenance of data used in outputs; the bill requires disclosures and explainability measures appropriate to risk level.

A requirement focused on providing affected individuals and administrative users with accessible, understandable information about the existence, purpose and effect of algorithmic analyses and the reasoning behind decisions to enable contestation and meaningful understanding.

The capability of an AI system to produce traceable, intelligible summaries or explanations of how inputs, model logic and data provenance contributed to a specific output, presented at a level appropriate to stakeholders and proportional to system impact.

Known as 'explicabilidad,' this principle mandates that AI systems' decision-making processes must be traceable, transparent, and their outcomes justifiable to human operators and citizens. This is vital for building trust, ensuring accountability, and allowing for effective review and challenge of AI-driven decisions.

The ability to describe the logic and functioning of an algorithm in plain language that is understandable to non-experts, ensuring meaningful transparency.

A foundational requirement for medical AI where the system must provide a human-understandable rationale for its outputs, essential for clinicians to justify treatment decisions.

The technical and procedural capacity to trace and understand the logic behind an AI system's output or decision-making process.

The degree to which a human can understand the cause of a decision or output of an AI system. It refers to the ability to provide clear reasoning for AI decisions, especially crucial for 'black box' models.

Explainability refers to the characteristic of an artificial intelligence system that allows a specified human audience to understand the rationale behind the system's actions, decisions, and outputs. This involves not just presenting results but also making the underlying processes and contributing factors transparent and comprehensible, fostering trust and accountability.

Within the context of IEEE 7001-2021, explainability refers to the ability of an autonomous system to provide explanations of its decisions and operations, tailored to the needs and technical literacy of different stakeholder groups, such as developers, operators, and end-users.

Implicitly addressed through the strategy's objectives related to ethical development and responsible deployment, suggesting that future regulations will require AI systems to offer understandable explanations for their actions, particularly when impacting individuals.