Artificial intelligence and the denial of privacy

Alongside availability, traceability and integrity, confidentiality is one of the four pillars of IT security. With the accelerating roll-out of artificial intelligence services to the general public, confidentiality is one of the pillars most under threat. A key principle of artificial intelligence technologies is algorithmic learning: by feeding data into the system, users contribute to the algorithms’ training datasets.

Best practice: do not submit a confidential document for translation by an algorithm without knowing what will happen to the source document afterwards. Do not upload third-party photographs to a tool without their consent. Do not ask an algorithm to process personal data without first ensuring that it complies with the GDPR.