How do we make AI trustworthy?
Due to the nature of AI-based systems, projects of this kind involve not only traditional product and project risks but also additional AI-specific risks, such as the AI-based system making incorrect decisions due to inaccurate or insufficiently representative training data. Furthermore, incorrectly selected training data can lead to the AI learning phenomena such as bias, overfitting, underfitting, or other quality issues, as well as making incorrect decisions. In addition, a certain proportion of people—in this specific case, a portion of users or those affected—may not accept decisions made by autonomous systems or even artificial intelligence.
It is therefore possible that an AI-based system may objectively meet all requirements or even produce significantly better results than conventional systems or human decision-makers, yet it may not be purchased or used because users lack trust in the system.
This is why new quality characteristics are emerging in the field of artificial intelligence, such as “trustworthiness”, which expresses the degree of trust a user has, or “explainability”, which provides AI users with a context-specific and understandable explanatory model, for example, to help them understand a decision that has been made. Depending on the context, other quality characteristics may also play a role in the use of AI-supported systems. Here, too, the relevant quality characteristics must be assessed, and their fulfillment must be ensured through appropriate quality assurance.
When selecting appropriate quality characteristics and metrics, testing them, and verifying compliance with standards and legal regulations, imbus can advise you and support you with appropriate QA measures and tests.
Ihr Ansprechpartner bei imbus
Mr. Tilo Linz