Share
August 1, 2024, marks a significant milestone as the European Union’s Artificial Intelligence (AI) Act officially comes into effect. This is the world’s first comprehensive legislation designed to regulate AI technology responsibly. So, what does this mean for businesses?
The EU AI Act is the result of extensive negotiations aimed at establishing a harmonized legal framework "for the development, the placing on the market, the putting into service, and the use of artificial intelligence systems" in the EU. The Act aims to ensure that AI systems are developed and used safely, transparently, and with respect for fundamental human rights. Non-compliance with the EU AI Act can result in financial penalties of up to €35 million or 7% of annual global revenue, whichever is higher.
AI systems are classified based on risk levels, from unacceptable risk (banned) to minimal risk (little or no regulation).
Safety and transparency: Ensure that AI does not pose risks to users and society.
Human Rights protection: Ensure AI does not violate fundamental rights.
Promote innovation: Support the development of advanced AI technologies within a clear legal framework.
The EU AI Act has a broad scope, covering most AI activities within the EU market. It defines AI systems as machine-based systems designed to operate with varying levels of autonomy, capable of adapting post-deployment and making decisions, predictions, or content based on input data to influence physical or virtual environments. This definition aligns with the OECD’s definition and applies to AI systems placed on the EU market or used within the EU. Entities subject to the Act include providers, deployers, importers, distributors, and manufacturers of AI products linked to the EU market. AI systems deployed outside the EU but whose outputs are used within the EU must also comply with the Act.
Thus, the Act will impact many businesses, especially those developing AI systems, but also those not primarily tech companies but deploying or merely using AI in certain cases. The impacts of the AI Act will be detailed below.
According to some assessments, the AI Act mainly targets large American tech companies, currently the primary builders and developers of the most advanced AI systems. However, other tech companies will also be significantly affected and closely monitored when operating in the EU market or using EU citizens' data. The AI Act presents both challenges and opportunities for businesses as follows:
Severe penalties: EU member states will enforce the regulation and ensure businesses comply. Non-compliant businesses can face hefty fines of up to €35 million or 7% of annual revenue for using prohibited AI systems. Non-compliance with other requirements, including GPAI, can incur fines up to €15 million or 3% of annual revenue. Providing incorrect, incomplete, or misleading information can result in fines up to €7.5 million or 1.5% of revenue.
Restrictions on AI in the workplace: The EU AI Act imposes strict limitations on using AI in the workplace, particularly in recruitment, employee monitoring, and performance evaluation. AI systems must ensure transparency, fairness, and privacy protection, requiring businesses to invest in safety and ethical measures. This can complicate processes, human evaluation, and performance reviews, increasing administrative workload.
Increased compliance costs: Businesses will need to invest in compliance with the new AI Act regulations. This includes costs for risk assessment, system upgrades, and employee training. Smaller businesses may find it more challenging to bear these costs compared to larger corporations. Risk assessments indicate that compliance with the new rules will cost companies around €6,000 to €7,000 for an average high-risk AI system worth €170,000 by 2025. Additional costs may arise from human oversight and verification.
Impact on global competitiveness: Global businesses will need to adjust their AI systems to comply with EU regulations, potentially leading to inconsistencies in AI standards worldwide. This could complicate international business operations and require flexible management.
Building trust with customers and the public: The AI Act will establish a legal framework that ensures trust in AI from customers and the public. Transparency and safety are prioritized, helping businesses build credibility and trust.
Promoting investment and innovation: The Act aims to promote investment and innovation in AI through mechanisms like "regulatory sandboxes" – controlled environments that allow new technologies to be tested under supervision. Each country will establish at least one national sandbox by June 2026, enabling businesses to test innovative solutions without immediate regulatory barriers.
Special support for SMEs and Start-ups: Special measures are in place to reduce the regulatory burden and support small and medium-sized enterprises (SMEs) and start-ups. Initiatives like Networks of AI Excellence Centres, the Public-Private Partnership on AI, Data and Robotics, Digital Innovation Hubs, and Testing and Experimentation Facilities will support companies in developing AI.
Participating in the community and sharing best practices: Providers of general-purpose AI models (GPAI) and the open-source community can participate in forums created by the AI Office to share best practices and contribute to codes of conduct and practice. This fosters a learning and continuous improvement community.
The EU AI Act is a significant step forward in responsibly regulating AI technology. Businesses must quickly adapt to and comply with the new regulations to seize opportunities and mitigate risks. Investing in AI safety and ethics not only ensures legal compliance but also builds credibility and trust with customers.
Share
