Introduction to Responsible AI : what security for AI?
This white paper is intended to serve as an introduction to the complex and fascinating world of AI security.
Aimed at tech professionals, be they experts in AI or not, this document seeks to provide a pragmatic overview of the security issues associated with these disruptive technologies. We will use the example of ChatGPT to illustrate the challenges, potential solutions and best practice for the secure use of these tools. However, obviously, AI security does not only concern ChatGPT.
We'll be delving into the intricacies of data confidentiality, ethics, the risks of adversarial attacks, and other crucial aspects of AI security. We'll also be looking at regulatory, technical, organisational and behavioural solutions using practical, relevant examples, in particular ChatGPT.