The rapid pace of technological development is leading to an enormous increase in the use of artificial intelligence (AI). The use of AI systems offers many opportunities, but there are also new risks.
Your contact
In particular, we focus on the question of assessing the reliability of the AI. Based on the IDW audit standard PS 861 for auditing AI systems, we support you in the quality assurance of your AI system.
Advantages of quality assurance for AI systems
At PKF Fasselt, we have in-depth expertise in financial and compliance matters. This knowledge is crucial to ensure that the requirements of an AI system comply with regulations. As external experts, we bring an objective perspective to the process and can act as a neutral party to ensure the quality of your AI system. This can strengthen the trust of customers, business partners and supervisory authorities. Risks associated with your AI system can be identified at an early stage and risk minimisation strategies developed so that financial risks and potential sanctions can be avoided.
Our audit is based on the following criteria
Ethical and legal requirements for AI
The criterion aims to organise the development, introduction and use of AI in such a way that it is in line with the company's ethically indispensable values as well as legal and regulatory requirements.
Traceability
The aim here is for the company to explain the decisions made by the AI system in a comprehensible manner. This includes the transparency and explainability of the data and AI algorithms or AI models used. Transparency refers to the provision of information, i.e. the data records and process steps that led to the decision must be documented in a comprehensible manner. Explainability refers to the ability to make both the data and models of an AI system and the decisions based on them understandable for the user.
IT security
AI systems must fulfil IT security requirements. These include confidentiality, integrity, availability, authorisation, authenticity and binding nature. In addition, special specifics of AI systems must be taken into account.
Performance capability
The AI system should be able to meet the requirements defined by the company based on internal operating practices or technology.