The EC proposes regulation and actions to foster trustworthy AI in Europe
Today, 21st of April, the Commission unveils the proposal for European artificial intelligence regulation. Combining the proposal with an updated Coordinated Plan on AI, the EC aims to guarantee the safety and fundamental rights of Europeans while strengthening AI uptake, investment and innovation across the EU.
The new Coordinated Plan on AI proposes joint actions with member states for ensuring all efforts are aligned with the European Strategy on AI and the European Green Deal. It also aims to speed up recovery from the COVID-19 pandemic with accelerated investments in AI. The plan utilizes Digital Europe and Horizon Europe programmes, the Recovery and Resiliency facility and Cohesion Policy programmes to allocate the funding. The plan includes a new public-private partnership and utilization of digital innovation hubs for fostering AI excellence. It also aims to create conditions for AI development through investment in critical computing capacities and data sharing, integrating trust into AI policies and building strategic leadership in high-impact sectors, such as environment, health and robotics.
The regulatory proposal categorizes AI systems by four different levels. The levels range from minimal to unacceptable risk. The package is focused on controlling "high-risk" applications before granting entry to the market. For example, applications related to employment, transportation and law enforcement are considered to contain high-risk. Those systems need to fill five obligations, assessing the quality of data, bias, explainability, human oversight, security and accuracy. High-risk systems are registered in an EU database, and re-assessment is required if the system changes substantially. Applications manipulating human behaviour to circumvent users' free will and social scoring by the governments are categorically banned, as they are considered to contain unacceptable risk. Biometric surveillance on public spaces is also prohibited, with some exceptions.
In addition to the risk-based approach, the legal framework seeks to reduce administrative and financial burdens for business. Targeting especially small and medium-sized enterprises (SMEs), it supports innovation with sandboxing schemes. The proposal also includes the creation of the "European Artificial Intelligence Board".
As industrial AI is at the focal point of BDVA/DAIRO community, we welcome the anticipated proposal. Having a direct impact on the data-driven artificial intelligence experimentation and innovation in Europe, the regulation carries great importance for the association and our members. The association has already started working on position paper through different task forces. Workshops shaping the paper are arranged during the course of May. If you are a member interested in contributing to the work or have any questions, send us an email to firstname.lastname@example.org.
Following a White Paper released in 2020 (you can find the BDVA/DAIRO response here), the objective of the regulation is to ensure a lawful AI respecting the fundamental rights of the EU. It is the first attempt to regulate artificial intelligence, setting a benchmark for rules governing the AI systems.