The term governance of foundation models is particularly at home in the fields of artificial intelligence, digital transformation, big data and smart data. Foundation models are very large artificial intelligence models that can be used for many different tasks, for example for text or image recognition.
In this context, governance means the rules, processes and controls that companies and organisations use to ensure that foundation models are used responsibly. This includes, for example, how the data from which the model learns is collected, who is authorised to use these models and how misuse can be prevented.
An illustrative example: Imagine a company uses a foundation model to automatically assess CVs. Without clear governance, biases in the model could lead to unfair decisions. With smart governance, guidelines are set on how the model is trained, tested and controlled so that it works as fairly and transparently as possible.
Governance of foundation models is essential to ensure that artificial intelligence can be used confidentially, securely and in the interests of the company and society.















