EU in the Age of AI: Five Things You Should Know
On April 19th, Reshaping Work organised an exclusive event in Brussels, convening companies, trade unions, think tanks, and policy makers to discuss the role of the EU in the age of artificial intelligence. Here are five key insights you should be aware of.
1. AI deployment in the workplace requires new protections
Throughout history, technological change has been a constant, continuously reshaping our economic landscape. The industrial revolutions serve as evidence that tech advancements do not necessarily lead to job losses on the aggregate; however, they do prompt shifts across sectors that can sometimes be quite “painful” for workers. As we stand on the brink of new changes, poised to affect various sectors and skill requirements, it is clear that such transformations will impact the labour market.
Mr. Andrea Glorioso, a policy officer at the European Commission’s Directorate-General for Employment, Social Affairs and Inclusion, reminds us that the history of industrial changes also highlighted the importance of social protection systems to cushion against economic shocks, which was and remains essential to reap the benefits of progress while ensuring social cohesion. Today, with AI driving similar transformations, the emphasis lies on creating robust protections. Such measures are vital to soften the impact of these changes and to safeguard those navigating a new technological revolution.
2. EU embraces risk-based approach in regulating AI
The European Union’s Artificial Intelligence Act (AI Act) is setting a global precedent in the regulation of AI technologies, differentiating between acceptable uses and those deemed too risky. The AI Act employs a risk-based framework categorizing AI systems into four main levels:
- Unacceptable Risk: At this level, certain AI practices, such as social scoring systems or indiscriminate mass data scraping, are flagged for their high risk. The AI Act takes a firm stance here, prohibiting such uses outright to defend users’ individual rights.
- High Risk: AI applications in sensitive areas like recruitment and healthcare come under closer scrutiny due to their profound potential impact. Deployment in these fields is conditional on passing stringent ethical and regulatory evaluations to ensure the highest ethical standards are upheld.
- Limited or Transparency Risk: This category includes AI systems that could potentially deceive, like highly realistic chatbots or deepfakes. While not outright harmful, they must operate with transparency, making it clear to users that they are engaging with AI, not humans.
- Minimal or No Risk: For AI applications assessed as largely harmless, the green light is given with little to no restraint. That said, there is a nudge for developers and users to embrace best practices through voluntary codes of conduct, fostering ethical usage.
Mr. Martin Ulbrich, a policy officer at the Directorate‑General for Communications Networks, Content and Technology at the European Commission, highlights that the AI Act entails a structured timeline for implementation, providing a phased approach that allows stakeholders ample time to adapt. Under this plan, Member States are required to phase out prohibited AI systems within 6 months of the AI Act coming into force. Rules pertaining to General Purpose AI (GPAI) will follow after 12 months, with a period of 24 to 36 months allocated for integrating regulations related to high-risk applications.
3. The impact of AI on fundamental rights needs to be accessed
The forthcoming AI Act calls for a Fundamental Rights Impact Assessment for AI systems classified as high-risk and deployed by or on behalf of public bodies and certain specific private operators, notably those in credit scoring or managing creditworthiness, as well as health and life insurance sectors, underscoring the EU’s dedication to upholding fundamental rights. Central to the Act, especially for AI systems outlined in Annex III, this assessment is vital. It delves deep into how and where the AI will be used, identifies which individuals and groups could be affected, and pinpoints the specific risks involved. The Act also mandates a blueprint for human oversight, ready to be activated if these risks materialize.
This rigorous approach underlines the EU’s commitment to ensuring AI is used responsibly and ethically, safeguarding citizens’ rights across the board, underlines Mr. Ulbrich.
4. The AI Office will be responsible for compliance and oversight
The EU AI Office, with some details yet to be clarified, is being set up as a central body within the European Commission with a wide range of responsibilities: from evaluations and fines to the oversight of general-purpose AI systems. According to Mr. Ulbrich, launching the AI Office requires the meticulous preparation of around 60 supporting documents to ensure smooth, standard-compliant operations.
Alongside the AI Office, the AI Act grants the national competent authorities the obligation to monitor high-risk AI applications and conduct market surveillance under Annex II. Member States are already preparing for this; Spain has established a new dedicated office; others are considering tasking their data protection office, cybersecurity agency or other pre-existing institutions.
At a broader level, the European Artificial Intelligence Board brings together high-level representatives from national authorities to offer advice and assistance to the Commission, and to play a supporting role in the implementation and enforcement of the Regulation. This is supplemented by the Advisory Forum, which draws participants from a cross-section of society — industry, start-ups, SMEs, civil society, and academia — ensuring a diverse range of insights and expertise regarding AI systems, and by a scientific expert panel. Together, these bodies form a comprehensive governance framework for AI, aiming to uphold the EU’s high standards for ethical AI usage.
5. The EU already has policy tools in place to mitigate AI risks
While there are ongoing discussions among many stakeholders about the possible need for new regulatory instruments focusing on the use of AI at the workplace, existing policy and regulatory instruments can already address some of the risks and challenges, points out Mr. Glorioso of the European Commission.
As a union with both economic and social objectives, the EU strategically employs active labour market policies, skilling initiatives, and measures to boost competitiveness in order to responsibly deploy AI in the world of work. These policies are designed to prepare the workforce for AI-induced changes, aiming to minimize disruption and prevent social unrest. This comprehensive approach underlines the importance of not only adapting technologically but also responding swiftly with appropriate policies, which are not exclusively regulatory in nature, to successfully integrate AI into the labour market, notes Mr. Glorioso.
Navigating the complex landscape of AI deployment in work and business operations requires a multistakeholder approach to policy development. By engaging a diverse group of stakeholders, including industry leaders, trade unions, start-ups, SMEs, civil society, and academia, Reshaping Work aims to build a consensus that values diverse perspectives and fosters an inclusive digital environment. This collaborative approach is crucial to ensure that AI technologies are developed and deployed ethically, transparently, and in alignment with the public interest. You can join us in pursuing this collaborative effort during the flagship conference on October 23 and 24 in Amsterdam, or the Reshaping Work Dialogue on AI scheduled for 2025.
#AIAct #EU #AI #LabourMarket #Workers #ResponsibleAI #FutureOfWork
Interested in this and similar stories? Subscribe to our bi-monthly newsletter to stay ahead of the curve with the latest updates and insights on the future of work: https://reshapingwork.net/contact/
