# Trustworthy and Ethical Assurance Platform Standard

###### **Version:** 0.0.1
###### **Last Modified:** 23rd April 2024

## 1. Introduction

This document presents a standard for the Trustworthy and Ethical Assurance (TEA) Platform, which facilitates building and evaluating assurance cases for data-driven technologies, particularly in the context of digital twins or artificial intelligence.

The standard defines the structure, relationships, and content of the core elements used to construct arguments about the trustworthiness and ethical implications of these technologies.

## 2. Description of Core Elements

We begin with a definition of each of the five core elements of an assurance case:

1. **Top-Level Goal Claim:** A statement asserting a desirable property or characteristic of the system or technology under consideration.
2. **Property Claim:** A statement that helps specify the top-level goal claim and defines a measurable requirement for the project or system under consideration.
3. **Strategy:** A course of action or approach that can help break the task of assuring a top-level goal claim into a set of related property claims.
4. **Evidence:** An artefact that justifies a linked property claim's validity and grounds an assurance case.
5. **Context:** Additional information that clarifies the scope or boundary conditions of a top-level goal claim.

### Relationships

The following relationships are considered valid links between the core elements defined above:

- **Goal Claim:** 
- Parents:
- `null`
- Children:
- `context` (many)
- `strategy` (many)
- `property claim` (many)
- **Property Claim:**
- Parents:
- `goal` (one)
- `strategy` (one)
- `property claim` (one)
- Children:
- `property claim` (many)
- `strategy` (many)
- `evidence` (many)
- **Strategy:**
- Parents:
- `goal` (one)
- `property claim` (one)
- Children:
- `property claim` (many)
- **Evidence:**
- Parents:
- `property claim` (many)
- Children:
- `null`
- **Context:**
- Parents:
- `goal` (one)
- Children:
- `null`

## 3. Element Details

This section provides additional information about the elements described above:

### 3.1 Top-Level Goal Claim

- **Name:** `goal_claim`
- **Fields**:
- Element Identifier (e.g `G1`):
- Goal Description: a statement asserting a desirable property or characteristic of the system or technology under consideration.
- **Example:** ==to draft==
- **Metadata:**
- Tags
- **Constraints:**
- Only one per assurance case.

### 3.2 Property Claims

- **Name:** `property_claim`
- **Element Identifier:** `P{1...n}`
- **Fields**:
- Property Claim Description
- **Description:** A statement that defines a measurable property for the project or system under consideration.
- **Example:** ==to draft==
- **Metadata:**
- Tags
- **Constraints:** 
- Can be nested to represent hierarchical relationships. Their identifiers should reflect this (e.g. P1 -> P1.1 -> P1.1.1)

### 3.3 Strategies

- **Name:** `strategy`
- **Element Identifier:** `S{1...n}`
- **Fields**:
- Strategy Description
- **Description:** A course of action or approach that can help break the task of assuring a top-level goal claim into a set of related property claims.
- **Example:** ==to draft==
- **Metadata:**
- Tags
- **Constraints:** 
- Does not include any content that affects the validity the argument.

### 3.4 Evidence

- **Name:** `evidence`
- **Element Identifier:** `E{1...n}`
- **Fields**:
- Element Description
- URI (e.g. DOI, URL)
- **Description:** An artefact that justifies a linked property claim's validity and grounds an assurance case.
- **Example:** ==to draft==
- **Metadata:**
- Tags
- Version of associated evidence document
- **Constraints:**
- Always the lowest level of an assurance case.

### 3.5 Context

- **Name:** `context`
- **Element Identifier:** `C{1...n}`
- **Description:** Additional information that clarifies the scope or boundary conditions of a top-level goal claim.
- **Example:** ==to draft==
- **Metadata:**
- Tags
- Version of associated evidence document
- **Constraints:** Must be relevant and provide meaningful clarification.

## 4. Additional Information

### Best Practices

- Use clear and concise language for all elements.
- Assurance cases should avoid arguments that require more than 3 levels of nested property claims.
- Where relevant evidence should adhere to the FAIR principles—findable, accessible, interoperable, reusable.

### Resources

- TEA Platform Documentation: [https://alan-turing-institute.github.io/AssurancePlatform/](https://alan-turing-institute.github.io/AssurancePlatform/)
- GSN Standard: [https://scsc.uk/gsn?page=gsn%202standard](https://scsc.uk/gsn?page=gsn%202standard)
- [Goal Structuring Notation Community Standard Version 3.](https://scsc.uk/r141C:1?t=1)
- FAIR Principles: [https://www.go-fair.org/fair-principles/](https://www.go-fair.org/fair-principles/)
- [TEA Platform - Legacy](https://assuranceplatform.azurewebsites.net)

## 5. Versioning

- **Version:** 0.0.1
- **Last Modified:** 23rd April 2024

This is a draft version of the TEA platform standard.

Future versions may include additional sections, elements, data types, or constraints based on platform development and user feedback.

## 6. Licensing

==Open license to be identified==