32 Agentic AI Safety Experts Focus Group. (2024). Guidelines for agentic AI safety: Volume 1. https://e-space. mmu.ac.uk/635454/1/Safer%2BAgentic%2BAI%2BFoundations%2BPublicationDraft%2BI1D1Jly24. docx%2B%5BIssue%2B2%5D%20%281%29.pdf 33 Royal Academy of Engineering, National Engineering Policy Centre. (2023). Towards autonomous systems in healthcare. https://nepc.raeng.org.uk/media/mmfbmnp0/towards-autonomous-systems-in-healthcare_-jul-2023-update.pdf 34 World Economic Forum. (2022). Earning digital trust: Decision-making for trustworthy technologies. https://www3. weforum.org/docs/WEF_Earning_Digital_Trust_2022.pdf 35 Agentic AI Safety Experts Focus Group. (2024). Guidelines for agentic AI safety: Volume 1. https://e-space. mmu.ac.uk/635454/1/Safer%2BAgentic%2BAI%2BFoundations%2BPublicationDraft%2BI1D1Jly24. docx%2B%5BIssue%2B2%5D%20%281%29.pdf 36 Chan, A., et al. (2024). Visibility into AI agents. https://arxiv.org/abs/2401.13138 37 Gabriel, I., et al. (2024). The ethics of advanced AI assistants. https://arxiv.org/abs/2404.16244 38 Ibid. 39 Ibid. 40 Ibid. 41 AI Seoul Summit (2024). International scientific report on the safety of advanced AI: Interim report. https://assets. publishing.service.gov.uk/media/6716673b96def6d27a4c9b24/international_scientific_report_on_the_safety_of_ advanced_ai_interim_report.pdf 42 Chan, A., et al. (2024), Visibility into AI agents. https://arxiv.org/abs/2401.13138 43 Royal Academy of Engineering/ National Engineering Policy Centre. (2023). Autonomous systems: A workshop on cross- cutting governance. https://nepc.raeng.org.uk/media/2hsh552k/as-workshop-report-v4.pdf 44 Ibid. 45 Chan, A., et al. (2024), Visibility into AI agents. https://arxiv.org/abs/2401.13138 46 AI Seoul Summit (2024). International scientific report on the safety of advanced AI: Interim report. https://assets. publishing.service.gov.uk/media/6716673b96def6d27a4c9b24/international_scientific_report_on_the_safety_of_ advanced_ai_interim_report.pdf 47 World Economic Forum. (2024). Advancing data equity: An action-oriented framework. https://www.weforum.org/ publications/advancing-data-equity-an-action-oriented-framework/ 48 Risk is defined as the combination of the probability of an occurrence of harm and the severity of that harm if it occurs. See: International Organization for Standardization/International Electrotechnical Commission. (2022). ISO/IEC 22989:2023: Information technology – artificial intelligence – artificial intelligence concepts and terminology. https://www. iso.org/standard/74296.html 49 Chan, A., et al. (2024), Visibility into AI agents. https://arxiv.org/abs/2401.13138 50 Royal Academy of Engineering/ National Engineering Policy Centre (2020). Safety and ethics of autonomous systems: Project overview. https://nepc.raeng.org.uk/media/nqnhktgq/nepc-safety-and-ethics-of-autonomous-systems.pdf 51 World Economic Forum. (2024). AI value alignment: Guiding artificial intelligence towards shared human goals. https:// www.weforum.org/publications/ai-value-alignment-guiding-artificial-intelligence-towards-shared-human-goals/ 52 Doshi-Velez, F., & Kim, B. (2017). Towards a rigorous science of interpretable machine learning. https://arxiv.org/ abs/1702.08608. 53 Chan, A., et al. (2024). Visibility into AI agents. https://arxiv.org/abs/2401.13138 54 Royal Academy of Engineering, National Engineering Policy Centre. (2023). Towards autonomous systems in healthcare. https://nepc.raeng.org.uk/media/mmfbmnp0/towards-autonomous-systems-in-healthcare_-jul-2023-update.pdf 55 World Economic Forum. (n.d.). Safe systems and technologies. AI Governance Alliance. Retrieved October 4, 2024, from https://initiatives.weforum.org/ai-governance-alliance/safesystems Navigating the AI Frontier: A Primer on the Evolution and Impact of AI Agents 27

Navigating the AI Frontier 2024