At Rise of AI 2024, Dr. Volker Wissing, Germany's Federal Minister for Digital Affairs and Transport, outlines the federal government's AI agenda in a short video address. He opens on AI's double edge in cybersecurity, useful for detecting vulnerabilities and predicting attacks but also usable to generate new attack methods faster than before. He then turns to regulation, describing the EU AI Act, expected to enter into force in July 2024, as an innovation-friendly compromise he helped shape, to be implemented with the lightest possible touch through standards developed with the European Committee for Standardization and input from businesses. He stresses that AI governance is a global issue, citing a G7 code of conduct for advanced AI and Germany's continued engagement at the AI Seoul Summit. He presents Germany's Mission AI initiative, launched in 2023, including new regional AI hubs starting in Kaiserslautern and Berlin, new AI quality and test standards, and events connecting AI startups with small and medium-sized enterprises, summarizing the government's approach as education, intelligent regulation, promotion and inspiration.
Key takeaways
- Wissing says AI's own strengths, fast method generation and near-unlimited learning ability, could also become its biggest cybersecurity risk.
- The EU AI Act, expected to enter into force in July 2024, will be implemented through standards developed with the European Committee for Standardization.
- The G7 agreed a common code of conduct requiring advanced AI to align with shared values and not harm democracies or support terrorism.
- Germany's Mission AI initiative is opening regional AI hubs, starting in Kaiserslautern and followed by Berlin, to advise startups on trustworthy AI.
- Mission AI includes developing transparent AI quality and test standards and events networking AI startups with small and medium-sized enterprises.
Detailed notes
AI's double edge in cybersecurity
Wissing opens by noting AI can identify vulnerabilities quickly, detect unfamiliar or suspicious activity that traditional systems might miss, respond to attacks in real time, and use machine learning on historical data to predict future attacks. He warns the same properties, the pace at which new attack methods can be generated and AI's virtually unlimited ability to learn, could also become its biggest risk, making it essential to train enough skilled people to use AI adeptly and to build trustworthiness into systems from the earliest development stage.
Implementing the EU AI Act with a light touch
Wissing states the federal government's stance is that AI must not enable discrimination or threaten democracy, requiring intelligent, forward-looking and practical regulation, which he says has been achieved in the EU AI Act, expected to enter into force in July 2024. He says finding innovation-friendly rules during the negotiations was a personal priority, and that implementation will now rely on standards, with the European Committee for Standardization developing them and businesses, sellers, buyers, users and regulators all invited to contribute their expertise.
Aligning AI governance internationally
Wissing argues AI is not only a European issue but a global one, requiring Germany's rules to align with international efforts. He points to a G7 common code of conduct stipulating that advanced AI must be developed in line with shared values and must not be used to harm democracies, manipulate people, or support terrorism or other criminal activity, describing it as establishing international guardrails to hold AI developers accountable. He says Germany will continue this exchange at the AI Seoul Summit, jointly hosted by South Korea and the UK in late May.
Mission AI: hubs, standards and startup networking
Wissing presents Mission AI, launched the previous year to keep Germany a pioneer in AI development and use. As part of it, the federal government is setting up regional AI hubs, the first opening in Kaiserslautern that summer to advise founders, startups and companies on the trustworthiness of their AI applications, let them trial and improve applications, and connect them to research, innovation and growth partners; a second hub in Berlin will focus on bringing AI's opportunities to life for citizens through exhibits. Mission AI also aims to develop transparent, widely applicable AI quality and test standards, work that began in December with key stakeholders, and is preparing a series of events, the first slated for July, to network AI startups with small and medium-sized enterprises. He summarizes the government's approach as education, intelligent regulation, promotion and inspiration, arguing that without AI the German and European economies will not remain competitive.
Quotes
AI is a powerful tool. It is therefore imperative that we are aware of its opportunities and threats.
Volker Wissing
Without AI the German and European economies will not be competitive.
Volker Wissing
Questions answered in this talk
When was the EU AI Act expected to enter into force?
According to Volker Wissing, Germany's Federal Minister for Digital Affairs and Transport, at Rise of AI 2024, the EU AI Act was expected to enter into force in July 2024, following negotiations in which he prioritized innovation-friendly rules.
What is Germany's Mission AI initiative?
According to Wissing, Mission AI, launched in 2023, includes regional AI hubs starting in Kaiserslautern and followed by Berlin, the development of transparent AI quality and test standards, and events networking AI startups with small and medium-sized enterprises.
What international AI commitments has Germany made?
According to Wissing, Germany helped the G7 agree a common code of conduct requiring advanced AI to align with shared values and not be used to harm democracies, manipulate people or support terrorism, and Germany continued this exchange at the AI Seoul Summit, co-hosted by South Korea and the UK in late May.