Safe and Ethical AI (SEA) Platform Network · Linking Artificial Intelligence Principles (LAIP)

1. Accountability and Transparency

...We have implemented audit and risk assessments to test our models as the baseline of our oversight methodologies....

Published by ADP

1. Transparency and Explainability

...• Ensuring traceability by building an audit trail to document the AI system development and decisionmaking process, implementing a black box recorder that captures all input data streams, or storing data appropriately to avoid degradation and alteration....

Published by ASEAN

1. Transparency and Explainability

...• Facilitating auditability by keeping a comprehensive record of data provenance, procurement, preprocessing, lineage, storage, and security....

Published by ASEAN

1. Transparency and Explainability

...Deployers should, however, note that auditability does not necessarily entail making certain confidential information about business models or intellectual property related to the AI system publicly available....

Published by ASEAN

1. Transparency and Explainability

...A risk based approach can be taken towards identifying the subset of AI enabled features in the AI system for which implemented auditability is necessary to align with regulatory requirements or industry practices....

Published by ASEAN

3. TRANSPARENCY

...The operation and impacts of an AI system should be transparent, traceable, auditable and generally explainable to a degree appropriate to its use and potential risk profile so outcomes can be understood and challenged, particularly where they relate to people....

Published by the Law, Society and Ethics Working Group of the AI Forum，New Zealand

Responsibility and accountability.

...Appropriate oversight, impact assessment, audit and due diligence mechanisms, including with regard to whistleblower protection, should be developed to ensure accountability for AI systems and their impact throughout their lifecycle....

Published by OFFICE OF THE CHIEF OF MINISTERS UNDERSECRETARY OF INFORMATION TECHNOLOGIES

Published by Beijing Academy of Artificial Intelligence (BAAI); Peking University; Tsinghua University; Institute of Automation, Chinese Academy of Sciences; Institute of Computing Technology, Chinese Academy of Sciences; Artifical Intelligence Industry Innovation Strategy Alliance (AITISA); etc.

Build a strong network and data security barrier.

...Support the Global Data Security Initiative, deepen cooperation between China and Africa in the fields of network and data security, jointly promote the research and application of network security technology and products, and create auditable, supervised, traceable, and trustworthy artificial intelligence technology....

Published by China Africa Internet Development and Cooperation Forum

6. We set the framework.

...They are transparent, auditable, fair, and fully documented....

Published by Deutsche Telekom

3. Traceable.

...DoD’s AI engineering discipline should be sufficiently advanced such that technical experts possess an appropriate understanding of the technology, development processes, and operational methods of its AI systems, including transparent and auditable methodologies, data sources, and design procedure and documentation....

Published by Defense Innovation Board (DIB), Department of Defense (DoD), United States

VII. Accountability

...auditability of AI systems is key in this regard, as the assessment of AI systems by internal and external auditors, and the availability of such evaluation reports, strongly contributes to the trustworthiness of the technology....

Published by European Commission

VII. Accountability

...External auditability should especially be ensured in applications affecting fundamental rights, including safety critical applications....

Published by European Commission

Auditability

... auditability...

Published by Fairness, Accountability, and Transparency in Machine Learning (FAT/ML)

· 8) Judicial Transparency

...Any involvement by an autonomous system in judicial decision making should provide a satisfactory explanation auditable by a competent human authority....

Published by Future of Life Institute (FLI), Beneficial AI 2017

· 5. The Principle of Explicability: “Operate transparently”

...Technological transparency implies that AI systems be auditable, comprehensible and intelligible by human beings at varying levels of comprehension and expertise....

Published by The European Commission’s High-Level Expert Group on Artificial Intelligence

2. Artificial intelligence should operate on principles of intelligibility and fairness.

...To ensure that our use of AI does not inadvertently prejudice the treatment of particular groups in society, we call for the Government to incentivise the development of new approaches to the auditing of datasets used in AI, and to encourage greater diversity in the training and recruitment of AI specialists....

Published by House of Lords of United Kingdom, Select Committee on Artificial Intelligence

2. Continued attention and vigilance, as well as accountability, for the potential effects and consequences of, artificial intelligence systems should be ensured, in particular by:

...a. promoting accountability of all relevant stakeholders to individuals, supervisory authorities and other third parties as appropriate, including through the realization of audit, continuous monitoring and impact assessment of artificial intelligence systems, and periodic review of oversight mechanisms;...

Published by 40th International Conference of Data Protection and Privacy Commissioners (ICDPPC)

3. Artificial intelligence systems transparency and intelligibility should be improved, with the objective of effective implementation, in particular by:

...c. making organizations’ practices more transparent, notably by promoting algorithmic transparency and the auditability of systems, while ensuring meaningfulness of the information provided, and...

Published by 40th International Conference of Data Protection and Privacy Commissioners (ICDPPC)

· Governance

...Access to global markets should be conditioned on domestic regulations meeting these global standards as determined by an international audit, effectively preventing development and deployment of systems that breach red lines....

Published by IDAIS (International Dialogues on AI Safety)

· Emergency Preparedness Agreements and Institutions

...To facilitate these agreements, we need an international body to bring together AI safety authorities, fostering dialogue and collaboration in the development and auditing of AI safety regulations across different jurisdictions....

Published by IDAIS (International Dialogues on AI Safety)

· Independent Global AI Safety and Verification Research

...Eventually, comprehensive verification could take place through several methods, including third party governance (e.g., independent audits), software (e.g., audit trails) and hardware (e.g., hardware enabled mechanisms on AI chips)....

Published by IDAIS (International Dialogues on AI Safety)

5. Accountability and timely correction

...Artificial intelligence solutions should be auditable and accountable....

Published by Megvii

5. Safety and Controllability

...The transparency, interpretability, reliability, and controllability of AI systems should be improved continuously to make the systems more traceable, trustworthy, and easier to audit and monitor....

Published by National Governance Committee for the New Generation Artificial Intelligence, China

Chapter 3. The Norms of Research and Development

...In the phases of algorithm design, implementation, and application, etc., improve transparency, interpretability, understandability, reliability, and controllability, enhance the resilience, adaptability, and the ability of anti interference of AI systems, and gradually realize verifiable, auditable, supervisable, traceable, predictable and trustworthy AI....

Published by National Governance Committee for the New Generation Artificial Intelligence, China

4. Accountable and responsible

...Algorithmic systems should also be regularly peer reviewed or audited to ensure that unwanted biases have not inadvertently crept in over time....

Published by Government of Ontario, Canada

3. Auditability

...auditability...

Published by Personal Data Protection Commission (PDPC), Singapore

Principle 6 – Transparency & Explainability

...The degree to which the system is traceable, auditable, transparent, and explainable is dependent on the context and purpose of the AI system and the severity of the outcomes that may result from the technology....

Published by SDAIA

· Build and Validate:

...This enables the identification of the AI decision and its respective analysis which facilitates its auditability as well as its explainability....

Published by SDAIA

Plan and Design:

...2 Organizations can put in place additional instruments such as impact assessments, risk mitigation frameworks, audit and due diligence mechanisms, redress, and disaster recovery plans....

Published by SDAIA

· Prepare Input Data:

...3 The documentation of the process is necessary for auditing and risk mitigation....

Published by SDAIA

· 1) Accountability:

...Artificial intelligence should be auditable and traceable....

Published by Youth Work Committee of Shanghai Computer Society

3. Organizations are accountable for the actions of AI systems, and should build systems that are auditable;

...Organizations are accountable for the actions of AI systems, and should build systems that are auditable;...

Published by The Treasury Board Secretariat of Canada (TBS)

· Algorithmic fairness

... Formulate guidelines and principles on solving bias and discrimination, potential mechanisms include algorithmic transparency, quality review, impact assessment, algorithmic audit, supervision and review, ethical board, etc....

Published by Tencent Research Institute

· "Black box" technology

... Promote algorithmic transparency and algorithmic audit, to achieve understandable and explainable AI systems...

Published by Tencent Research Institute

Third principle: Understanding

...While the ‘black box’ nature of some machine learning systems means that they are difficult to fully explain, we must be able to audit either the systems or their outputs to a level that satisfies those who are duly and formally responsible and accountable....

Published by The Ministry of Defence (MOD), United Kingdom

Responsibility and accountability

...United Nations system organizations should have in place appropriate oversight, impact assessment, audit and due diligence mechanisms, including whistle blowers’ protection, to ensure accountability for the impacts of the use of AI systems throughout their lifecycle....

Published by United Nations System Chief Executives Board for Coordination

· Responsibility and accountability

...Appropriate oversight, impact assessment, audit and due diligence mechanisms, including whistle blowers’ protection, should be developed to ensure accountability for AI systems and their impact throughout their life cycle....

Published by The United Nations Educational, Scientific and Cultural Organization (UNESCO)

· Responsibility and accountability

...Both technical and institutional designs should ensure auditability and traceability of (the working of) AI systems in particular to address any conflicts with human rights norms and standards and threats to environmental and ecosystem well being....

Published by The United Nations Educational, Scientific and Cultural Organization (UNESCO)

3. Traceable

...The department's AI capabilities will be developed and deployed such that relevant personnel possess an appropriate understanding of the technology, development processes and operational methods applicable to AI capabilities, including with transparent and auditable methodologies, data sources and design procedures and documentation....

Published by Department of Defense (DoD), United States

6. Auditability

...auditability...

Published by ACM US Public Policy Council (USACM)

6. Auditability

...Models, algorithms, data, and decisions should be recorded so that they can be audited in cases where harm is suspected....

Published by ACM US Public Policy Council (USACM)

3 Ensure transparency, explainability and intelligibility

...It must be possible to audit an AI technology, including if something goes wrong....

Published by World Health Organization (WHO)

3 Ensure transparency, explainability and intelligibility

...Health care institutions, health systems and public health agencies should regularly publish information about how decisions have been made for adoption of an AI technology and how the technology will be evaluated periodically, its uses, its known limitations and the role of decision making, which can facilitate external auditing and oversight....

Published by World Health Organization (WHO)

1. Accountability and Transparency

...We have implemented audit and risk assessments to test our models as the baseline of our oversight methodologies....

1. Transparency and Explainability

...• Ensuring traceability by building an audit trail to document the AI system development and decisionmaking process, implementing a black box recorder that captures all input data streams, or storing data appropriately to avoid degradation and alteration....

1. Transparency and Explainability

...• Facilitating auditability by keeping a comprehensive record of data provenance, procurement, preprocessing, lineage, storage, and security....

1. Transparency and Explainability

...Deployers should, however, note that auditability does not necessarily entail making certain confidential information about business models or intellectual property related to the AI system publicly available....

1. Transparency and Explainability

...A risk based approach can be taken towards identifying the subset of AI enabled features in the AI system for which implemented auditability is necessary to align with regulatory requirements or industry practices....

Accountability

...Accountability By Design: All AI systems must be designed to facilitate end to end answerability and auditability....

3. TRANSPARENCY

...The operation and impacts of an AI system should be transparent, traceable, auditable and generally explainable to a degree appropriate to its use and potential risk profile so outcomes can be understood and challenged, particularly where they relate to people....

Responsibility and accountability.

...Appropriate oversight, impact assessment, audit and due diligence mechanisms, including with regard to whistleblower protection, should be developed to ensure accountability for AI systems and their impact throughout their lifecycle....

· Be Ethical

...This may include, but not limited to: making the system as fair as possible, reducing possible discrimination and biases, improving its transparency, explainability and predictability, and making the system more traceable, auditable and accountable....

Build a strong network and data security barrier.

6. We set the framework.

...They are transparent, auditable, fair, and fully documented....

3. Traceable.

VII. Accountability

...auditability of AI systems is key in this regard, as the assessment of AI systems by internal and external auditors, and the availability of such evaluation reports, strongly contributes to the trustworthiness of the technology....

VII. Accountability

...External auditability should especially be ensured in applications affecting fundamental rights, including safety critical applications....

Auditability

... auditability...

· 8) Judicial Transparency

...Any involvement by an autonomous system in judicial decision making should provide a satisfactory explanation auditable by a competent human authority....

· 5. The Principle of Explicability: “Operate transparently”

...Technological transparency implies that AI systems be auditable, comprehensible and intelligible by human beings at varying levels of comprehension and expertise....

2. Artificial intelligence should operate on principles of intelligibility and fairness.

2. Continued attention and vigilance, as well as accountability, for the potential effects and consequences of, artificial intelligence systems should be ensured, in particular by:

3. Artificial intelligence systems transparency and intelligibility should be improved, with the objective of effective implementation, in particular by:

...c. making organizations’ practices more transparent, notably by promoting algorithmic transparency and the auditability of systems, while ensuring meaningfulness of the information provided, and...

· Governance

...Access to global markets should be conditioned on domestic regulations meeting these global standards as determined by an international audit, effectively preventing development and deployment of systems that breach red lines....

· Emergency Preparedness Agreements and Institutions

...To facilitate these agreements, we need an international body to bring together AI safety authorities, fostering dialogue and collaboration in the development and auditing of AI safety regulations across different jurisdictions....

· Independent Global AI Safety and Verification Research

...Eventually, comprehensive verification could take place through several methods, including third party governance (e.g., independent audits), software (e.g., audit trails) and hardware (e.g., hardware enabled mechanisms on AI chips)....

5. Accountability and timely correction

...Artificial intelligence solutions should be auditable and accountable....

5. Safety and Controllability

...The transparency, interpretability, reliability, and controllability of AI systems should be improved continuously to make the systems more traceable, trustworthy, and easier to audit and monitor....

Chapter 3. The Norms of Research and Development

4. Accountable and responsible

...Algorithmic systems should also be regularly peer reviewed or audited to ensure that unwanted biases have not inadvertently crept in over time....

3. Auditability

...auditability...

Principle 6 – Transparency & Explainability

...The degree to which the system is traceable, auditable, transparent, and explainable is dependent on the context and purpose of the AI system and the severity of the outcomes that may result from the technology....

· Build and Validate:

...This enables the identification of the AI decision and its respective analysis which facilitates its auditability as well as its explainability....

Plan and Design:

...2 Organizations can put in place additional instruments such as impact assessments, risk mitigation frameworks, audit and due diligence mechanisms, redress, and disaster recovery plans....

· Prepare Input Data:

...3 The documentation of the process is necessary for auditing and risk mitigation....

· 1) Accountability:

...Artificial intelligence should be auditable and traceable....

3. Organizations are accountable for the actions of AI systems, and should build systems that are auditable;

...Organizations are accountable for the actions of AI systems, and should build systems that are auditable;...

· Algorithmic fairness

... Formulate guidelines and principles on solving bias and discrimination, potential mechanisms include algorithmic transparency, quality review, impact assessment, algorithmic audit, supervision and review, ethical board, etc....

· "Black box" technology

... Promote algorithmic transparency and algorithmic audit, to achieve understandable and explainable AI systems...

Third principle: Understanding

...While the ‘black box’ nature of some machine learning systems means that they are difficult to fully explain, we must be able to audit either the systems or their outputs to a level that satisfies those who are duly and formally responsible and accountable....

Responsibility and accountability

...United Nations system organizations should have in place appropriate oversight, impact assessment, audit and due diligence mechanisms, including whistle blowers’ protection, to ensure accountability for the impacts of the use of AI systems throughout their lifecycle....

· Responsibility and accountability

...Appropriate oversight, impact assessment, audit and due diligence mechanisms, including whistle blowers’ protection, should be developed to ensure accountability for AI systems and their impact throughout their life cycle....

· Responsibility and accountability

...Both technical and institutional designs should ensure auditability and traceability of (the working of) AI systems in particular to address any conflicts with human rights norms and standards and threats to environmental and ecosystem well being....

3. Traceable

6. Auditability

...auditability...

6. Auditability

...Models, algorithms, data, and decisions should be recorded so that they can be audited in cases where harm is suspected....