With the finalization of the AI Act on the horizon, many questions, viewpoints, proposals, and discussions are arising regarding its effective implementation. To achieve AI innovation that is truly trustworthy we need the conditions that enable the processes and practices embodying the values and vision that the act represents. The “coding” of law into practice needs constructive approaches and mechanisms to ensure that we get it right from the beginning.  

In the framework of the 4th AI Alliance Assembly in Madrid, the Panel “Implementing the AI Act in practice” brings together multiple views and experiences that will delveinto the details and key questions that are arising alongside the AI act implementation. .  

Several initiatives are already underway in anticipation of the AI Act's adoption.  These initiatives include AI standardization, regulatory sandboxes and supervisory activities. AI standardization is essential for the interoperability andtechnical implementation of requirements of the act. For example, different standardisation deliverables have here been requested by the Commission, in areas such as risk management systems, governance and quality of datasets, record-keeping, transparency and information, human oversight, accuracy, data quality and governance, cybersecurity, quality management systems, and conformity Regulatory sandboxes are controlled spaces where companies can test new AI products and services in a real-world environment, but under managed conditions. This allows companies to experiment with different approaches and to identify potential risks and challenges. Regulatory sandboxes can for example play an important role in facilitating the implementation of the AI Act by helping businesses to understand and comply with the new regulations. Likewise, supervisory authorities will play a vital role in ensuring that the AI Act is implemented and enforced effectively. In fact, we need to understand how and in which form enforcement of the act can be realised with different forms of certification and market surveillance mechanisms.   

The panel session will take place on November 17, 2023, from 11:45 AM to 1:00 PM, at Casa de América in Madrid, Spain. It aims at shedding light on these ongoing efforts and fostering discussion and exchange between the various initiatives. I look forward to moderating this session that will feature speakers bringing insights and expertise from different sectors.  Huub Janssen, representing the Dutch Authority on Digital Infrastructure, Ursula Pachl, Deputy Director General of BEUC,  Kilian Gross, Head of AI Policy Development and Coordination Unit at DG Connect, Emilia Tantar, Head of Delegation for Luxembourg on AI standardization, Corinna Schulze, Director EU Government Affairs, SAP and Manuel Marina Breysse, CEO, Cardiologist & Co-founder, IDOVEN This panel will explore cooperation among national competent authoritiesand the promotion of learning and synergies. With this, visions on the roadmap of standardisation at the EU level and the challenges for compliance will be addressed, including viewpoints from the different speakers on the opportunities, challenges, and concerns of adhering to the AI Act. The role of consumers in shaping AI Act implementation and the importance of standards for informed consumer choices will be also taken into account.A deep dive into the initiatives led by the European Commission to promote the practical implementation of the AI Act will be offered by speakers, who will discuss topics such as establishing communication channels between industry stakeholders and national supervisory authorities, strategies for ensuring that AI standards remain up-to-date and effective, and the barriers faced by companies seeking to innovate in AI technology.  

The discussion will provide insights into the practical implementation of initiatives aiming at making AI trustworthy, secure, and aligned with EU regulations. I invite you to join us and gain insights into how these different initiatives are coming together to implement the AI Act.