At a Rise of AI Virtual Chat in 2020, Thomas Heilmann, member of the German Bundestag, discusses his book NEUSTAAT with host Fabian Westerheide. Heilmann describes NEUSTAAT as more than 100 concrete reform proposals for digitizing and modernizing German government, developed over a year and a half of research by a working group that traveled to China and South Korea to study what makes societies more innovative. He argues the German government proved surprisingly fast and effective during the Corona crisis when focused on a single topic, and that the pandemic forced schools, courts and parliament to adopt digital tools that existing law already allowed but no one had executed. He proposes a state-backed retirement fund invested globally by an independent agency, university investment funds modeled on MIT, and European Investment Bank backing for startup IPOs to address Europe's lack of late-stage funding and exits. He also outlines a "future lobby" to represent not-yet-existing industries in lawmaking, data auditors who would certify how companies use personal data, and a risk-based approach to AI regulation, arguing that clear rules, not the absence of regulation, help businesses invest with confidence.
Key takeaways
- Heilmann's working group researched innovative states in China and South Korea before writing NEUSTAAT.
- Three NEUSTAAT proposals were already folded into Germany's Corona economic stimulus package.
- He proposes the European Investment Bank anchor-invest in European startup IPOs to fix weak exit options.
- He wants licensed "data auditors" to certify how companies use personal data instead of minimizing data collection.
- He argues AI regulation should be risk-based and clear, since uncertainty deters startup investment more than strict rules.
Detailed notes
NEUSTAAT: a reform agenda from a year of research
Heilmann explains that the NEUSTAAT project began a year and a half before the talk, when a working group set out to find how to bring more innovation into Germany and Europe. After nearly a year of research, including trips to China and South Korea, the group concluded that a well-functioning state is a key condition for innovation and set modernization as a central task for the Bundestag session starting in 2021, with a discussion period through the end of 2020 and execution from late 2021. The book covers education, startup funding, the retirement system and administrative efficiency, and proposes over 100 ideas, including an annual 32-billion-euro public retirement fund, a "future agency," a blockchain-based euro, school curricula covering economy, IT and social behavior, and a unified app store for public services. Heilmann compares the ambition to the Prussian reforms after Napoleon's conquests, arguing similarly radical change is needed now.
Corona as a forced digitization test
Heilmann says the pandemic moved most political meetings online and forced an enormous volume of laws to be processed under time pressure, run partly from his home office. He notes the German government reacted faster than expected, and argues the system works well in a crisis focused on one topic but is slow when many issues compete for attention. As an example, 2.9 million students moved to online study almost overnight, something that six months earlier would have been called a ten-year project; many teachers had no computer before the crisis. Three NEUSTAAT proposals, he says, are already part of the government's Corona stimulus package: modernizing public registers, a notification tool, and an online public-services gazette.
Funding startups and a new retirement fund
Heilmann attributes Europe's startup funding gap partly to the near-absence of pension-fund capital, unlike the US where pension funds are the largest source of venture money, and to weak exit options beyond acquisition by Google, Facebook, Amazon or Microsoft. He proposes that the European Investment Bank anchor-invest in European tech IPOs as a long-term minority shareholder for roughly five years, and that universities run co-investment funds alongside private investors, modeled on MIT, to improve the transfer of research into companies. On pensions, he proposes an independent, Bundesbank-observed agency investing part of income globally with only minority, non-board stakes, citing a failed early-20th-century German pension fund that lost its assets after investing in World War One bonds as a cautionary precedent.
A "future lobby" and skepticism on Gaia-X
Heilmann argues that existing companies are skilled at lobbying and getting heard in parliamentary hearings, while businesses and technologies that do not yet exist have no one to represent them. He cites e-fuels as an example: because no company producing them existed when an environmental law was written, e-fuels' CO2-neutral properties are not counted in current balance sheets. He proposes a foundation-style "future lobby" to advocate for not-yet-existing industries during lawmaking. On Gaia-X, he supports the goal of a European cloud standard with stronger data protection and lower cost than AWS, but says it is currently "a little over-engineered," with too many participants who lack technical understanding, risking delay.
Data sovereignty through data auditors
Heilmann says Germany is at the beginning of digitalization, not the end, and expects data volumes to keep exploding as computing power grows. He argues against Germany's traditional principle of minimizing data collection ("Datensparsamkeit"), proposing instead data sovereignty: companies that profile personal data would need certification from a licensed "data auditor," modeled on financial auditors, who has full access to a company's data use and can revoke the license for misuse. He argues this system would have to apply to Facebook, Google, Amazon and Alibaba if they want to operate in Europe, and separately describes an earlier conversation about a public data marketplace where audited companies could exchange data rather than one company owning it alone.
Risk-based AI regulation and early standards
Heilmann argues AI needs regulation proportional to potential harm, for example stricter rules for AI controlling car brakes than for lower-risk logistics systems, with the same data-auditor structure assessing risk. Asked how startups can absorb the compliance burden, he says "the answer can't be no regulation, the answer must be clear regulation," because clarity, even if strict, lets founders plan and raise investment, unlike uncertainty. He draws a parallel to Germany's early 220-volt electrical standard and argues Europe should set AI standards early, citing the GSM mobile standard as a precedent, warning that without an early standard a single company could build a monopoly. In closing, he expects the reforms to shift power and wealth, compares public resistance to the backlash Ludwig Erhard faced over 1948 price liberalization, and says his personal goal for 2021 is to see the modernization agenda adopted by the governing coalition, regardless of his own role in it.
Quotes
The answer can't be no regulation, the answer must be clear regulation.
Thomas Heilmann
We have to realize that we are at the beginning of the digitalization, not at the end.
Thomas Heilmann
One of the key conditions is that you have a well-functioning state.
Thomas Heilmann
Questions answered in this talk
What is NEUSTAAT and what does it propose?
According to Thomas Heilmann at a Rise of AI Virtual Chat in 2020, NEUSTAAT is his book proposing over 100 reforms to modernize the German state, covering education, startup funding, a public retirement fund, administrative digitization and infrastructure.
Why did German schools and government move online so quickly during Corona?
Thomas Heilmann explains that the legal basis for remote schooling and digital government processes already existed but had never been executed; the crisis forced 2.9 million students and much of parliament online almost overnight, something previously estimated to take a decade.
How does Thomas Heilmann propose regulating AI in Europe?
Heilmann proposes a risk-based approach where the potential damage of an AI application determines the level of regulation, assessed by licensed data auditors, arguing that clear rules help businesses invest even if the rules are strict.