This paper provides key guidance to ensure the development, deployment, and use of AI systems align with seven crucial requirements. These requirements encompass human agency and oversight, technical robustness and safety, privacy and data governance, transparency, non-discrimination and fairness, environmental and societal wellbeing, and accountability. The guidelines emphasise the use of both technical and non-technical methods for implementing these requirements.

To uphold ethical principles, the guidelines advocate for the development, deployment, and use of AI systems that respect human autonomy, prevent harm, ensure fairness, and promote explicability. Specific technical standards for AI systems, ranging from data processing to technological safety, are recommended. Additionally, stakeholder involvement is highlighted, urging collaboration in both the development and use of AI systems, with a particular focus on addressing biases.

The report highlights beneficial opportunities and critical concerns associated with AI. Opportunities include potential improvements in transport systems, healthcare, education, and environmental sustainability. However, critical concerns involve AI’s impact on human rights, potential conflicts with existing regulations, and its use in autonomous warfare.

To operationalise the key requirements of trustworthy AI, the report introduces the Trustworthy AI Assessment List. This list is designed to be concrete and non-exhaustive, tailored to the specific use case and context of an AI system. Governance for trustworthy AI is also addressed, suggesting the involvement of internal and/or external ethical experts or boards to identify potential conflicts and propose solutions. Furthermore, meaningful consultation with stakeholders at risk affected by an AI system is recommended, ensuring their involvement in relevant policymaking processes.

The report provides clear recommendations for ensuring trustworthy AI. It encourages adherence to the guidelines, the use of the Trustworthy AI Assessment List during development, deployment, or usage of AI systems, and the engagement of stakeholders to ensure diverse perspectives. The development of technical standards for AI systems, encompassing data processing to technological safety specifications, is also recommended. Continuous monitoring of AI systems throughout their lifecycle is stressed to guarantee ongoing compliance with Trustworthy AI principles.

In summary, the Ethics Guidelines for Trustworthy AI by the European Commission’s AI High-Level Expert Group lay out comprehensive recommendations to guide the responsible development, deployment, and use of AI systems. The guidelines cover crucial aspects such as human oversight, technical robustness, fairness, and accountability. The report acknowledges both the positive opportunities and critical concerns associated with AI and introduces practical tools like the Trustworthy AI Assessment List to operationalise these guidelines. Stakeholder involvement and continuous monitoring are key aspects emphasised for the sustained development of trustworthy AI systems.