The European Parliament has addressed the European Commission’s draft proposal for a new Artificial Intelligence Act, which was released in April. The new Parliament report comes courtesy of the agency’s Policy Department for Citizens’ Rights and Constitutional Affairs, and suggests that the law does not do enough to ensure the ethical use of AI-powered public identification systems.
With that in mind, the report goes on to argue for the inclusion of several amendments that would more strictly limit the use of AI technologies. Most notably, Parliament asked legislators to ban the use of surveillance systems in private homes and workspaces, and called for periodic reviews of the list of banned practices as new technologies emerge. It also pushed for a new section that provides more detail about those restricted applications.
In that regard, the report expressed particular concern about remote identification systems that can identify members of the public in real time. That includes biometric identification systems that use technologies like face or gait recognition. The original Commission draft prevents the police from using a real time surveillance network, but does allow law enforcement and border agents to use facial recognition to investigate serious crimes.
The Parliament report seems to ask the Commission to close some of those loopholes, or to at least provide more clarity about what is and is not allowed. In doing so, the report calls for greater transparency about how biometric systems are used, how conclusions are reached, and what kinds of technologies are admissible in court. Those provisions could prevent organizations from using biometric inferences while making certain kinds of decisions.
Parliament also wants the Commission to provide clearer definitions for biometric inferences and biometric data. Finally, the report requested new guidelines for automated consent management procedures to make it easier (and safer) for people to opt into biometric applications and services should they so wish to do so.
The broader goal is to ensure that any AI systems that do get used in Europe do not infringe on the civil liberties of the European public. Parliament backed the Commission’s efforts to regulate AI technologies, but felt that more rigor was needed to achieve those ends. As a result, Parliament’s report echoes statements from the European Data Protection Board and the European Data Protection Supervisor, both of which have argued that the Commission’s original draft proposal is too lax on the subject of facial recognition.
September 23, 2021 – by Eric Weiss