Monitoring & Retention
We keep the predictions of your AI systems in a vault, check them regularly for exposure towards conformity risks, and provide guidance towards mitigation.
How does this contribute to AI conformity?
We understand it’s not your core business to consistently keep an eye on your AI systems. We perform regular checks on the predictions from your employed AI systems to recognize pitfalls such as discrimination, false information and prejudice. We record these, report these to you and work with your project manager(s) to mitigate these issues.
For increased accountability and responsibility, keeping a paper trail of your predictions is important. This way, problems can be backtraced instead of ending up pointing fingers. Depending on your situation, local AI regulations might even require you to keep such paper trails for up to 6 months. We take care of this for you, in compliance with local privacy laws.
How does it work?
We will implement a setup that is as little invasive as possible, to export your prediction logs to our storage vault. We aim for this integration to take up as little time as possible from your engineers. Based on the nature of your organization, we curate a set of checks and our consultants will perform these checks periodically. Depending on your operation, you might be more interested in or more exposed towards certain types of risks. We will check across all types, but we prioritize the ones most important to your business.
From these checks, our consultants may encounter issues that need to be addressed. We will report these issues (as well as the findings in general) in our monthly reports towards your internal stakeholders. Additionally, we will discuss the encountered issues with the relevant project manager within your organization, to convert these into concrete tasks. Progress on these mitigation tasks will also be reported towards your internal stakeholders, and in part to the public.