POSITION | DIGITAL POLICY | ARTIFICIAL INTELLIGENCE
International Digital Policies on AI
Complementary international frameworks for harmonized and industry friendly technology standardisation.
31/05/2024
Executive Summary
International cooperation on digital policies becomes increasingly significant with the rise of the highly complex Artificial Intelligence (AI) systems. Internationally operating businesses face difficulties in compliance as soon as competing frameworks and legislation enforce contradicting regulation. The Federation of German Industries (BDI) calls for and will support international cooperation and planning certainty. Business friendly, unbureaucratic, harmonized, and trustworthy legal frameworks with actually risk-based and technology-neutral approaches to AI are vital for a thriving, competitive, globally connected and innovative European industry sector.
Key messages
German industry calls on legislators to address the following measures in the process of developing future-proof and balanced international AI governance.
▪ Ensure harmonized legislation with complementary frameworks rather than competing regulation to safeguard a consistent level playing field, enable innovation, and mitigate risks across markets.
▪ Maintain a topen regulation that does not give preference to any particular technology. Minimizing the negative impacts for society and environment while maximizing the development and application of innovative AI products and AI based solutions, is only possible by keeping to a risk-based AI regulation approach. Keep legislation proportionate and non-bureaucratic to minimize administrative burdens and foster innovation - the reporting obligations for companies should be kept to a minimum.
▪ Align definitions, sub-categories, terminologies, taxonomies, and ontologies on already existing wording.
▪ Promote AI awareness by reinforcing AI literacy, skilled labour and holistic approaches to the whole AI ecosystems, including the strengthening of semiconductor supply chains and HPC-programs.
▪ Promote awareness on cybersecurity for AI (on data, training, models) as well as the opportunities of AI for cybersecurity.
▪ Develop international, technical-sound, performance-based standards for an inclusive ethical use of AI, promoting democratic values, and providing legal certainties to European companies. Strengthening European participation and representation in international standardization bodies.
Bundesverband
Global Leadership for harmonized AI policy frameworks
In the course of rapid technological developments, multilateral forums surrounding the safe and trustworthy use of AI have simultaneously increased, the AI Safety Summit in Bletchley Park, the G7 Hiroshima Process and the AI Advisory Body by the United Nations being the largest frameworks on Artificial Intelligence regulation. These bodies established exchange formats and published reports and guidelines on the use of AI. The strong involvement of the US and the UK on the matter indicates their interests in taking up global leadership in international digital politics. The European Union also needs to develop an influential role in international AI governance structures. Cooperation, innovation-driven leadership and strong participation of state and non-governmental actors are key factors for an AI regulatory framework that enables European enterprises to compete on the international stage.
Robust regulation with common and clear definitions
If there is contradicting law, internationally operating businesses struggle with legal uncertainties and barriers to trade. Sometimes, these companies must individually adjust their product configurations and business solutions for every single country. Sometimes, this creates overwhelming hurdles for companies, especially for SMEs and start-ups. As a result, high-potential companies are withdrawing from lucrative markets. To prevent fragmentation and to establish a common rulebook, German industry recommends aligning AI definitions with already existing ones. Legislators should apply the same reasoning for establishing sub-categories of AI and related terminology, such as trustworthy, (high) risk, harm, bias, and risk management.
Additionally, BDI urges policymakers to avoid duplication of AI regulation (i.e. with sectoral legislation) and thereby prevent the associated legal uncertainties and double burdens for the economy. AI policies should enhance – not hinder - AI-based innovative businesses models and improve global market competition. In the swiftly changing digital environment, legislation must remain technological-neutral. While an innovation-open and value-based level playing field may foster AI products and applications from and in Europe, generalised regulation of technologies as such weakens particularly industrial research and creates a competitive disadvantage. A technology-neutral legislation, thus, is robust and safeguards upcoming technological developments in the first place, while still maintaining risk mitigation.
Promoting AI awareness and holistic AI ecosystems
Non-Bureaucratic and innovation friendly agreements
AI Regulation must stay proportionate and non-bureaucratic, so that companies of all sizes and especially small and medium sized companies as well as startups can realistically implement policies without excessive or unrealistic efforts. Reporting obligations are an important competitive factor for companies in the global environment. The BDI is calling for these reporting obligations to be minimised in order to maintain competitiveness. BDI calls on the policymakers to give reliable guidance, e.g., by setting clear and appropriate deadlines and instructions for implementation. To ensure its practicability, BDI urges the legislator to protect multistakeholder cooperation and to regularly consult relevant actors, especially the industry sector, in the legislative process. Similarly, practical applications such as regulatory sandboxes should remain.
Business-friendly regulation fosters innovation and investment, facilitating the economic and societal benefits new technologies have to offer. A strong securitization of AI, as seen in the AI summit or the G7 process, needs to be accompanied by measures giving incentives for innovation, such as education programmes and improvements to digital infrastructure. German industry appreciates if governments
adjust risk classifications on AI to their industry suitability and proportionality while fostering AI literacy, the usage of AI in administration and education and access to data and computing infrastructure.
Cherishing
societal and economic benefits
Key innovations, such as new technological developments and generative AI, and target computing devices benefit industry and society alike. AI applications can be used to effectively address urgent challenges such as the green transformation. Trustworthiness in these advancements is key for a successful digital age. Therefore, it is about time to improve society´s know-how and capabilities around AI, its potential for the industry sector, its positive impact on the labour market, and its general benefits. BDI urges to increase the flow of talent in the field of AI, tackle the shortage of skilled labour, and support diversity in the digital sector.
Viewing AI innovations in holistic ecosystems
Successful and competitive AI business models are highly dependent on a reliable digital policy framework. Alongside innovation-friendly regulation, companies need legally secure access to data, reliable access to high-performance computing infrastructure and the cutting-edge microchips on the market. The BDI is therefore calling for and will support collaboration and the strengthening of data spaces, HPC-programs, and the strengthening of semiconductor production and supply chains. The prevention of cyber risks in AI systems as well as the potential of AI to increase cyber security should be considered.
Determining the international market: setting standards for the use of AI
International Standardisation: strengthening European participation
As we aim to support democratic values, human rights, and environmental and social wellbeing, we also need to determine prevailing, international standards. At the same time, businesses need clear and common standards on the use of AI as they mitigate legal uncertainties, protect their competitiveness, enable global market access and large-scale production. The supporting effect of standards in the application of regulations has been successfully practised for over thirty years in the regulations on CE marking, which regulate the requirements for placing products on the market. This legislative principle is laid down in the New Legislative Framework (NLF). Regulations on AI with product-related requirements should therefore follow this NLF legislative principle. If Europe is able to introduce and enforce our standards on the use of AI, businesses get legal certainties and, following that, emerge stronger from the standardisation process. The global application of our standards thus also becomes a competitive advantage.
BDI urges and will support international standardisation bodies to coordinate the standardisation processes on AI adequately and calls for and will support the active participation of German and European stakeholders in international bodies. During the standardisation process, administrators must identify usability security and quality conditions for a safe and trustworthy use of AI together with relevant stakeholders from economy, academia, and civil society. A comprehensive representation of the stakeholders relevant to standardisation includes industry, science, associations and NGOs, government bodies and national certifiers. They need to adopt proportionate, market-driven, and consensus-based standards for a possible and realistic compliance, also for small and medium sized enterprises and startups.
Existing national and multilateral standardisation frameworks should be integrated into international formats so that companies only have to comply with one set of standards. European participation in the international standardisation bodies ISO/IEC should remain constant or grow even with increasing participation from other countries in order to maintain proportions. If other regions expand their position in the world of standardisation and thereby also consolidate their impact on the development of future technologies such as AI, this could lead to a significant competitive disadvantage for European companies. European representatives should also be present in foreign national standardisation bodies such as NIST. We recommend strengthening ETSI as a European standardisation organisation, which also has international significance, in order to ensure good coordination and cooperation between the various players in standardisation.
Imprint
Bundesverband der Deutschen Industrie e.V. (BDI)
Breite Straße 29, 10178 Berlin www.bdi.eu
T: +49 30 2028-0
Editor
Jacqueline Welsch Trainee
T: +49 30 2028-0000
j.welsch@bdi.eu
Polina Khubbeeva
Senior Manager Digitalization and Innovation
T: +49 30 2028-1586
p.khubbeeva@bdi.eu
BDI document number: D1898