The coverage paper, AI Assessments: Enhancing Confidence in AI, argues that efficient evaluations of AI techniques may help organisations deploy expertise that’s protected, dependable and compliant.
The report comes as international AI adoption accelerates and incidents linked to AI techniques proceed to rise — OECD knowledge reveals antagonistic AI incidents grew virtually twentyfold between November 2022 and January 2025.
Three varieties of AI assessments
The steering categorises AI evaluations into three varieties:
- Governance assessments, which look at whether or not inside buildings and insurance policies adequately handle AI dangers and reliability.
- Conformity assessments, which decide compliance with legal guidelines, rules and requirements.
- Efficiency assessments, which measure AI outputs in opposition to predefined high quality and efficiency standards.
EY and ACCA stress that these assessments, whether or not voluntary or obligatory, are important to company governance, danger administration and public confidence.
Nevertheless, present frameworks differ extensively by jurisdiction, methodology and scope, creating inconsistency and gaps in assurance.
Challenges and proposals
The report identifies challenges undermining present AI assessments, together with unclear aims, ambiguous terminology, a scarcity of certified suppliers and the speedy tempo of AI growth outstripping technical requirements.
To deal with these, the paper requires clearer definitions, standardised methodologies, and better skilled accountability.
Marie-Laure Delarue, EY International Vice-Chair, Assurance, mentioned:
“AI has been advancing quicker than many people might have imagined, and it now faces an inflection level, presenting unbelievable alternatives in addition to complexities and dangers. Rigorous assessments are an vital instrument to assist construct confidence within the expertise, and confidence is the important thing to unlocking AI’s full potential as a driver of development and prosperity.”
Helen Model, ACCA Chief Government, mentioned:
“As AI scales throughout the economic system, the power to belief the expertise is important for the general public curiosity. That is an space the place we have to bridge expertise gaps and construct belief within the AI ecosystem as a part of driving sustainable enterprise.”
International coverage momentum
The paper notes that policymakers in practically 70 nations have proposed greater than 1,000 AI coverage initiatives, with 204 enacted into regulation, in accordance with Stanford College.
Frameworks such because the EU AI Act, the UK AI Assurance Toolkit, and the US AI Motion Plan are driving regulatory exercise, however requirements and necessities stay fragmented.
EY and ACCA advocate that policymakers clearly outline the aim and standards of assessments, align requirements internationally to cut back prices, and construct capability to supply high-quality evaluations.
For enterprise leaders, the report urges consideration of voluntary assessments to reinforce governance and stakeholder confidence even within the absence of regulatory mandates.
