Artificial Intelligence (AI) confidence gets a boost through fresh guidelines issued by ACCA and EY

Artificial Intelligence (AI) confidence gets a boost through fresh guidelines issued by ACCA and EY

The Association of Chartered Certified Accountants (ACCA) and global professional services firm EY have jointly released a policy paper titled AI Assessments: Enhancing Confidence in AI. The report underscores the importance of trust in AI for the public interest, emphasizing the role of rigorous assessments in unlocking AI's full potential and managing its risks.

The key types of AI assessments identified in the paper include governance assessments, conformity assessments, and performance assessments.

Governance assessments evaluate whether appropriate corporate governance policies, processes, and personnel are in place to manage AI systems effectively. These assessments address risks, suitability, and reliability.

Conformity assessments focus on determining whether AI systems comply with relevant laws, regulations, standards, or policy requirements.

Performance assessments measure the quality of an AI system’s core functions, such as accuracy, non-discrimination, and reliability, often using quantitative metrics.

These assessments may be performed separately or combined and are essential for corporate governance, legal compliance, and ensuring trustworthy AI outcomes. The paper highlights their role in increasing confidence, effectiveness, and trust in AI technologies amidst accelerating adoption and evolving regulatory environments.

Policymakers are urged to back standardized, internationally aligned frameworks for AI assessments. The report addresses the challenges in the AI assessment field, such as inconsistent methodologies and a lack of skilled providers.

To bolster corporate governance and stakeholder trust, business leaders are encouraged to use both voluntary and mandated assessments. The paper offers new guidance to help businesses evaluate and trust their Artificial Intelligence (AI) systems. However, it does not specify a timeline for the implementation or adoption of these recommendations.

Marie-Laure Delarue, Global AI Leader at EY, emphasizes the critical role of these assessments in unlocking AI's full potential and managing its risks. Helen Brand, Chief Executive of ACCA, underscores the necessity of building trust in AI for the public interest.

[1] ACCA and EY, AI Assessments: Enhancing Confidence in AI, 2021. [2] Delarue, M-L., & Brand, H. (2021). Building Trust in AI: A Joint Policy Paper by ACCA and EY. Retrieved from https://www.ey.com/en_gl/ai/ai-assessments-enhancing-confidence-in-ai [3] ACCA. (2021). AI Assessments: Enhancing Confidence in AI. Retrieved from https://www.accaglobal.com/gb/en/professional-insights/insights/technology/artificial-intelligence/ai-assessments-enhancing-confidence-in-ai.html [4] EY. (2021). AI Assessments: Enhancing Confidence in AI. Retrieved from https://www.ey.com/en_gl/ai/ai-assessments-enhancing-confidence-in-ai [5] ACCA and EY. (2021). AI Assessments: Enhancing Confidence in AI. Retrieved from https://www.accaglobal.com/content/dam/ACCA_Global/professional-insights/technology/artificial-intelligence/ai-assessments-enhancing-confidence-in-ai.pdf

1. The joint policy paper by ACCA and EY, titled AI Assessments: Enhancing Confidence in AI, underscores the importance of governance assessments, conformity assessments, and performance assessments for managing risks and unlocking AI's full potential.
2. Governance assessments, as per the report, evaluate corporate governance policies, processes, and personnel to ensure effective management of AI systems, addressing risks, suitability, and reliability.
3. Performance assessments, focused on measuring an AI system's core functions such as accuracy, non-discrimination, and reliability, are essential for ensuring trustworthy AI outcomes and increasing confidence in AI technologies.
4. To bolster corporate governance and stakeholder trust, the report encourages businesses to use both voluntary and mandated assessments, offering new guidance on evaluating AI systems leveraging technology like artificial-intelligence, but does not specify a timeline for their implementation or adoption.

Read also: