Kai Zenner, Digital Policy Adviser at the European Parliament kicked off the first day at Identity Week Europe 2023 with an expert address delivered to a packed main keynote theatre.
Kai outlined the first set of rules on the AI Act. After 11 months of negotiations between member states and the EU Parliament, the next stage to implement the AI Act will be to find an agreement on the requirements for technology providers between member states and parliamentarians.
In the beginning, law enforcement authorities and member states pushed back on prohibitions proposed in the AI Act, and while this has progressed the
Delivering the first keynote of Identity Week Europe 2024, he set out the first set of rules for the AI Act, expanding on requirements.
“With a black and white approach, it becomes difficult to navigate”, he said.
“We cannot exclude positive use cases”, “this shows us what should not be prohibited in the future”. As the AI Act has been accepted, the useful purposes of biometric technology to enhance the quality of police investigations in catching offenders – and sharing biometric data with relevant authorities, such as border control bodies managing immigration – has been recognised as well as the “inherent” level of bias within technologies forgotten as solution providers mitigate bias.
While careful in their approach to begin with, the Act has been better defined through parliamentary votes to ensure that AI encompasses the potential for “surveillance” or anti-privacy practices with biometric and emerging technologies. Different political camps emerged for and against embracing AI technologies or creating supportive regulation. With a second wave of AI, deep learning and machine learning, the Act is needed to create standards surrounding the use of these technologies – minding privacy and obtaining consent for the collection, storage and exchange of personal AI data.
With machine regulation and GDPR, lots of legislation existed.
In his speech, Kai stressed the European Parliament’s next objectives to create AI legislation, including optimising government mechanisms and ensuring that EU level can work effectively with national level to observe rules for AI.
An issue with passing the regulation concerned overlapping competence and responsibilities with one party giving a statement that the other conflicted, while some considered there is too much legislation and not enough convergence.