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

E. Governability:

AI applications will be developed and used according to their intended functions and will allow for: appropriate human machine interaction; the ability to detect and avoid unintended consequences; and the ability to take steps, such as disengagement or deactivation of systems, when such systems demonstrate unintended behaviour.

Principle: NATO Principles of Responsible Use of Artificial Intelligence in Defence, Oct 22, 2021

Published by The North Atlantic Treaty Organization (NATO)

Related Principles

5. Governable.

DoD AI systems should be designed and engineered to fulfill their intended function while possessing the ability to detect and avoid unintended harm or disruption, and for human or automated disengagement or deactivation of deployed systems that demonstrate unintended escalatory or other behavior.

Published by Defense Innovation Board (DIB), Department of Defense (DoD), United States in AI Ethics Principles for DoD, Oct 31, 2019

4. Principle of safety

Developers should take it into consideration that AI systems will not harm the life, body, or property of users or third parties through actuators or other devices. [Comment] AI systems which are supposed to be subject to this principle are such ones that might harm the life, body, or property of users or third parties through actuators or other devices. It is encouraged that developers refer to relevant international standards and pay attention to the followings, with particular consideration of the possibility that outputs or programs might change as a result of learning or other methods of AI systems: ● To make efforts to conduct verification and validation in advance in order to assess and mitigate the risks related to the safety of the AI systems. ● To make efforts to implement measures, throughout the development stage of AI systems to the extent possible in light of the characteristics of the technologies to be adopted, to contribute to the intrinsic safety (reduction of essential risk factors such as kinetic energy of actuators) and the functional safety (mitigation of risks by operation of additional control devices such as automatic braking) when AI systems work with actuators or other devices. And ● To make efforts to explain the designers’ intent of AI systems and the reasons for it to stakeholders such as users, when developing AI systems to be used for making judgments regarding the safety of life, body, or property of users and third parties (for example, such judgments that prioritizes life, body, property to be protected at the time of an accident of a robot equipped with AI).

Published by Ministry of Internal Affairs and Communications (MIC), the Government of Japan in AI R&D Principles, Jul 28, 2017

· 2. RESPONSIBILITY MUST BE FULLY ACKNOWLEDGED WHEN CREATING AND USING AI

2.1. Risk based approach. The degree of attention paid to ethical AI issues and the nature of the relevant actions of AI Actors should be proportional to the assessment of the level of risk posed by specific AI technologies and systems for the interests of individuals and society. Risk level assessment shall take into account both known and possible risks, whereby the probability level of threats, as well as their possible scale in the short and long term shall be considered. Making decisions in the field of AI use that significantly affect society and the state should be accompanied by a scientifically verified, interdisciplinary forecast of socio economic consequences and risks and examination of possible changes in the paradigm of value and cultural development of the society. Development and use of an AI systems risk assessment methodology are encouraged in pursuance of this Code. 2.2. Responsible attitude. AI Actors should responsibly treat: • issues related to the influence of AI systems on society and citizens at every stage of the AI systems’ life cycle, i.a. on privacy, ethical, safe and responsible use of personal data; • the nature, degree and extent of damage that may result from the use of AI technologies and systems; • the selection and use of hardware and software utilized in different life cycles of AI systems. At the same time, the responsibility of AI Actors should correspond with the nature, degree and extent of damage that may occur as a result of the use of AI technologies and systems. The role in the life cycle of the AI system, as well as the degree of possible and real influence of a particular AI Actor on causing damage and its extent, should also be taken into account. 2.3. Precautions. When the activities of AI Actors can lead to morally unacceptable consequences for individuals and society, which can be reasonably predicted by the relevant AI Actor, the latter, should take measures to prohibit or limit the occurrence of such consequences. AI Actors shall use the provisions of this Code, including the mechanisms specified in Section 2, to assess the moral unacceptability of such consequences and discuss possible preventive measures. 2.4. No harm. AI Actors should not allow the use of AI technologies for the purpose of causing harm to human life and or health, the property of citizens and legal entities and the environment. Any use, including the design, development, testing, integration or operation of an AI system capable of purposefully causing harm to the environment, human life and or health, the property of citizens and legal entities, is prohibited. 2.5. Identification of AI in communication with a human. AI Actors are encouraged to ensure that users are duly informed of their interactions with AI systems when it affects human rights and critical areas of people’s lives and to ensure that such interaction can be terminated at the request of the user. 2.6. Data security. AI Actors must comply with the national legislation in the field of personal data and secrets protected by law when using AI systems; ensure the security and protection of personal data processed by AI systems or by AI Actors in order to develop and improve the AI systems; develop and integrate innovative methods to counter unauthorized access to personal data by third parties and use high quality and representative datasets obtained without breaking the law from reliable sources. 2.7. Information security. AI Actors should ensure the maximum possible protection from unauthorized interference of third parties in the operation of AI systems; integrate adequate information security technologies, i.a. use internal mechanisms designed to protect the AI system from unauthorized interventions and inform users and developers about such interventions; as well as promote the informing of users about the rules of information security during the use of AI systems. 2.8. Voluntary certification and Code compliance. AI Actors may implement voluntary certification systems to assess the compliance of developed AI technologies with the standards established by the national legislation and this Code. AI Actors may create voluntary certification and labeling systems for AI systems to indicate that these systems have passed voluntary certification procedures and confirm quality standards. 2.9. Control of the recursive self improvement of AI systems. AI Actors are encouraged to cooperate in identifying and verifying information about ways and forms of design of so called universal ("general") AI systems and prevention of possible threats they carry. The issues concerning the use of "general" AI technologies should be under the control of the state.

Published by AI Alliance Russia in AI Ethics Code (revised version), Oct 21, 2022 (unconfirmed)

Fourth principle: Bias and Harm Mitigation

Those responsible for AI enabled systems must proactively mitigate the risk of unexpected or unintended biases or harms resulting from these systems, whether through their original rollout, or as they learn, change or are redeployed. AI enabled systems offer significant benefits for Defence. However, the use of AI enabled systems may also cause harms (beyond those already accepted under existing ethical and legal frameworks) to those using them or affected by their deployment. These may range from harms caused by a lack of suitable privacy for personal data, to unintended military harms due to system unpredictability. Such harms may change over time as systems learn and evolve, or as they are deployed beyond their original setting. Of particular concern is the risk of discriminatory outcomes resulting from algorithmic bias or skewed data sets. Defence must ensure that its AI enabled systems do not result in unfair bias or discrimination, in line with the MOD’s ongoing strategies for diversity and inclusion. A principle of bias and harm mitigation requires the assessment and, wherever possible, the mitigation of these biases or harms. This includes addressing bias in algorithmic decision making, carefully curating and managing datasets, setting safeguards and performance thresholds throughout the system lifecycle, managing environmental effects, and applying strict development criteria for new systems, or existing systems being applied to a new context.

Published by The Ministry of Defence (MOD), United Kingdom in Ethical Principles for AI in Defence, Jun 15, 2022

5. Governable

The department will design and engineer AI capabilities to fulfill their intended functions while possessing the ability to detect and avoid unintended consequences, and the ability to disengage or deactivate deployed systems that demonstrate unintended behavior.

Published by Department of Defense (DoD), United States in DoD's AI ethical principles, Feb 24, 2020