Identity and AI: Questions of data security, trust and control | Computer Weekly

▤ Summary

AI-driven identity solutions can improve verification and reduce friction but introduce substantial compliance, privacy and ethical challenges because they rely on extensive personal data such as biometrics, behavioural and device signals. Organizations must treat identity decisions as governance matters, addressing lawful basis, necessity, retention and oversight rather than viewing these tools as mere technical controls. Privacy risks include surveillance creep and over-collection, while biased training data and opaque models can produce discriminatory or unfair access outcomes, so explainability and human review are essential. Adopt privacy-by-design, impact assessments, transparent notices and clear accountability to govern identity AI from the outset and avoid creating opaque, over-engineered risks.

▦ Linked Companies

Adobe

$1B+