This paper examines FRT as the first use case under NITI Aayog’s RAI principles and aims to establish a framework for responsible and safe development and deployment of FRT within India.
- FRT is an AI system which allows identification or verification of a person based on certain images or video data using complex algorithms.
Working of FRT
- FRT primarily seeks to accomplish three functions -
- Facial detection which relies on algorithms to detect presence of human face.
- Facial extraction which uses mathematical representations to identify distinctive features on individual faces.
- Facial recognition which involves automatic cross-referencing of a person’s facial features with pre-existing database.
Applications of FRT
- Security Related: Law and order enforcement (surveillance, identification of persons of interest, monitoring of crowd, screening for violation of public norms).
- Non-Security related:
- Ease of access in services (e.g. contactless onboarding at airports through Digi Yatra).
- Ease in usability such as unique IDs in educational institutions etc.
- Authentication for access to products, services, and public benefits.
Risks with FRT systems
- Design-based risks: Automation bias, discrimination, lack of accountability, misidentification/inaccuracy due to under-representations in databases.
- Rights-based issues: Privacy and lack of consent, informational autonomy, and processing of sensitive personal data etc.
Recommendations for responsible use of FRT
|
