Embedded Files
Welcome to the website for Sherpa Recommendations
What is the problem to be solved?
What is the problem to be solved?
Standards describe requirements and characteristics for products, methods and services. When ethics and human rights are not taken into consideration for all AI-related standards, there is a risk of serious negative implications for society and a reduction of trust in AI.
Who should act?
Who should act?
Standardisation committees, which should include ethicists, innovators and researchers, should work together to improve the standards to comply to ethical principles and human rights.
What should be done?
What should be done?
Standardization committees to bring ethics and human rights into the work on AI and convince the stakeholders about its necessity. Committees should adjust the requirements to comply with the ethical principles and human rights.
Key considerations
Key considerations
There are numerous ethics guidelines on AI that standardisation committees could draw on, such as:
Jobin et al (2019) provides an overview.
Examples of ways of implementing or furthering this recommendation include:
including it in corporate social responsibility annual reports,
including it in GDPR-rules
including it in ISO 9001 (though he does not like that option himself),
including it in an ethics and compliance programme
It is hard to reach an agreement on what ethical AI is, which hampers standard development.
SHERPA contribution
SHERPA contribution
SHERPA has produced a set of guidelines for operationalisation of ethics by design for developers and for users of smart information systems that can inform standardisation.
SHERPA has contributed to the standard on health and wellness apps that includes an annex on ethics.
SHERPA members are represented in national mirror committees of ISO/IEC JTC 1/SC 42 Artificial intelligence.
Page updated
Google Sites
Report abuse