On 14 June 2023, the European Parliament adopted its negotiating position on the Artificial Intelligence (AI) Act. The European Parliament's vote on the AI Act proposal marks a significant milestone toward the regulation of AI within the European Union, as it sets the baseline for inter-institutional negotiations, as further discussed below.
The proposed AI Act follows a risk-based approach, banning AI
applications that pose an unacceptable level of risk and imposing a
strict regime for high-risk use cases. It also establishes
obligations for AI providers and those deploying AI systems that
are tailored to the level of risk posed by the AI.
The European Parliament expanded its initial list of AI systems
with an unacceptable level of risk to include bans on intrusive and
discriminatory uses of AI – such as real-time remote
biometric identification systems in publicly accessible spaces,
biometric categorization systems using sensitive characteristics
(e.g., gender, race, ethnicity, citizenship status, religion and/or
political orientation), predictive policing systems (based on
profiling, location or past criminal behavior), emotion recognition
systems in law enforcement, border management, the workplace and
educational institutions, or untargeted scraping of facial images
from the internet or closed-circuit television (CCTV) footage to
create facial recognition databases.
Providers of 'foundation models' – i.e., AI models
trained on a large amount of unlabeled data that can be adapted to
many applications – will have to assess and mitigate possible
risks to health, safety or fundamental rights and register their
models in an EU database before releasing them on the EU market. In
addition to this, it will be mandatory for generative AI systems
based on such models to comply with a series of transparency
requirements, including an obligation to disclose when content is
AI-generated and to ensure safeguards against generating deep fake
content.
On the other hand, the European Parliament's position adds
exemptions for research activities and AI components provided under
open-source licenses, as well as the establishment in every EU
member state of at least one regulatory sandbox – a
controlled environment established by a public authority that
facilitates the safe development, testing and validation of
innovative AI systems for a limited time before their placement on
the market or putting them into service pursuant to a specific plan
under regulatory supervision.
As far as citizens' rights are concerned, the European
Parliament's position recognizes the right to file complaints
about AI systems and receive explanations of decisions that are
made based on high-risk AI systems that significantly impact
citizens' fundamental rights.
Following the European Parliament's adoption of its position
on the AI Act, the inter-institutional negotiations with the
European Council, European governments' representatives and the
European Commission in the so-called trilogue will take place. The
negotiations will kick off on 21 June, with a deal expected to be
reached by November 2023.
Expected to enter into force no later than the beginning of 2024,
the AI Act will not only have an extraterritorial reach – as
it will be applicable to providers and users outside the EU when
the output produced by the system is used within the EU – but
it also foresees fines of the higher of up to 40 million euros or
up to 7% of a company's total worldwide annual turnover for the
preceding financial year.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.