Key Considerations for Organisations Ahead of the EU AI Act

As artificial intelligence (AI) continues to evolve, organisations must navigate a landscape filled with both opportunities and challenges. With the upcoming EU AI Act coming into effect in February 2025, it is more important than ever for organisations to implement AI in a responsible and compliant manner. Keep reading to discover the key insights and considerations that organisations should focus on to ensure their AI initiatives align with regulatory requirements and ethical standards.

AI systems present unique governance challenges that differ significantly from traditional IT systems. These technologies require new approaches to ensure they are transparent and accountable. Three critical areas that organisations should focus on include:

· Robustness: AI systems must perform reliably under varying conditions. This involves rigorous testing to ensure that the system can handle diverse scenarios without failure or unintended consequences.

· Fairness: It is important to ensure that AI decisions are unbiased and equitable. Organisations need to actively work to identify and mitigate any biases in their AI models to prevent unfair or discriminatory outcomes.

· Interpretability: Unlike traditional systems, AI often requires the ability to explain its decisions clearly. This interpretability is essential for accountability, as stakeholders need to understand how decisions are made and be able to trust the AI’s outputs.

By addressing these governance challenges, organisations can mitigate potential risks and ensure that their AI systems function as intended.

Data quality is the backbone of effective AI systems. For AI to deliver accurate and reliable results, organisations must have a well-developed data strategy. This involves more than just collecting and managing data—it requires a comprehensive approach to data governance.

High-quality data supports the accuracy of AI outputs, giving organisations a competitive advantage. Additionally, strong data governance frameworks are crucial in ensuring that AI systems comply with regulatory standards, such as those outlined in the EU AI Act.

The rise of “shadow AI”—the unauthorised or unregulated use of AI by employees—poses a significant challenge for organisations. This can lead to inconsistencies in AI applications and potential regulatory breaches.

To address this, organisations should establish clear guidelines and support structures for AI implementation. By doing so, they can ensure that all AI initiatives are aligned with organisational objectives and ethical standards. Proactively managing shadow AI helps maintain control over AI usage and prevents unintended risks.

AI is becoming an integral part of various organisational functions, requiring existing roles to adapt and incorporate AI-specific knowledge. Key roles such as digital strategists, business developers, IT support staff, data supervisors, and information security coordinators must understand how AI differs from traditional IT systems.

These roles need to be prepared to manage the unique challenges that AI brings, from ensuring data quality to overseeing the ethical implications of AI decisions. By integrating AI knowledge across these roles, organisations can better manage the complexities of AI technologies.

Procurement is a critical factor in the successful and responsible implementation of AI. Organisations need to be informed and assertive in setting requirements during the procurement process. AI solutions often require customisation based on local data, meaning they cannot be entirely outsourced or standardised.

Maintaining control over the implementation and governance of AI systems is essential. Organisations must ensure that their AI solutions meet specific requirements and adhere to ethical standards, which can only be achieved through careful and knowledgeable procurement practices.

AI encompasses a wide array of technologies, from Large Language Models (LLMs) to more established algorithms. Understanding these differences is crucial because each type of AI technology may require different governance tools and approaches.

By distinguishing between the various AI technologies, organisations can apply the appropriate governance measures to manage each effectively. This nuanced approach helps ensure that AI technologies are used responsibly and in line with regulatory expectations.

AI implementation varies widely across organisations, from small-scale in-house pilots to the deployment of off-the-shelf products. Each approach requires a different considerations and organisational setup.

Tailoring the AI implementation process to the specific needs and context of the organisation is essential. Whether the AI initiative is a large-scale deployment or a smaller exploratory project, the implementation strategy must be customised to ensure success and compliance.

Fostering AI literacy across your organisation is critical for informed decision-making and effective AI implementation. As AI becomes more embedded in organisational processes, employees at all levels must understand the basics of AI and its implications.

Increasing AI literacy ensures that organisations can make informed decisions, invest in the right technologies, and fully leverage the potential of AI. This broad understanding also supports the responsible use of AI, helping to align its deployment with both ethical standards and regulatory requirements.

As the EU AI Act approaches, organisations need to prioritise responsible AI implementation. By focusing on key areas such as transparency, data quality, and proactive management, they can navigate the complexities of AI while ensuring compliance with emerging regulations. These insights and guidelines will be invaluable in helping organisations deploy AI technologies responsibly, ethically, and effectively, ensuring that their use aligns with both public values and regulatory standards.

Interested in learning more about how ADC can help your organisation prepare for the EU AI Act? Our team of experts are available for a no-obligation discussion. Feel free to reach out to Per-Erik Nyström (Senior Manager).