Reducing bureaucracy costs by 25% by 2029, offering more public services online and using in government and in courts are some of the projects Chancellor ‘s cabinet agreed to this past week.
Among the plans for integrating AI into everyday use is an online platform for export-based businesses to bundle information on export regulations and credit options, and expediting visa processing by using AI to review an applicant’s documents.
The “Modernization Agenda,” as Digitalization Minister has called it, comes on the heels of the similarly titled “High-Tech Agenda.” Presented by Dorothee Bär, federal minister of research, technology and space, in July, the “High-Tech Agenda” highlights AI as one of the six key technologies the government will promote, alongside biotechnology and microelectronics.
AI minister makes his debut
A more visible example of the government’s recent is the so-called Weimatar, an AI avatar of State Minister for Culture and Media . The Weimatar can speak 100 languages and is promised to be an example of “fair AI, that protects creativity and strengthens our democratic public sphere.” Its purpose is not only to more easily reach a broader audience on social media, but also to save time internally at the ministry with, for example, training videos that can be made quickly and kept up to date.
Holger Hoos, a Humboldt Foundation Professor of AI at RWTH Aachen University, does not believe the tech is understood well enough by those in power to promise a fair and responsible use of AI.
In 2023, a group of Humboldt Professors of AI, Germany’s preeminent researchers in the field, met with policymakers to prepare a list of suggestions for ethical and useful AI in government. This included the recommendation that the government be principally guided by science in its decision-making, not businesses.
“A government has a moral duty to follow competent advice. You can’t just follow the advice from industry, which often has vested interests in certain outcomes,” Hoos said.
At the same time, the expert in machine learning takes it as a good sign that Wildberger and the digitalization ministry have a core mission for improving Germany’s lagging digital infrastructure.
For his part, Wildberger has pointed out that Germany risks becoming if it does not embrace AI more swiftly.
“AI is the key to future growth, and AI is in full swing. It will change the world more — and is already doing so — than any technology before it,” he told the Funke media group. “If we don’t use AI widely, we risk more jobs than if we use it bravely and responsibly.”
AI is also widely used at the state and local levels. In the southern state of , for example, a raft of offers have been made public: from AI-text streamlining to a government chatbot in the city of Ludwigsburg. The state government also touts its use of , to “optimize public administration.”
Generative AI hype already waning
Questions are also being raised, both in traditional and social media, about whether Merz’s government is buying into the generative AI hype at the cost of more practical applications. Tech policy news site Netzpolitik accused the government of continuing Germany’s reputation for being behind the times in the digital world. “The government is thus once again succumbing” to the generative AI hype, just as it “subsides elsewhere,” they wrote.
Indeed, companies and governments worldwide have already bumped up against the limitations of the likes of and DALL-E, AI generators which provide over-generalized and sometimes inaccurate information. Experts like Hoos believe that if there are to be positive outcomes for citizens from using AI in government, energy should be directed instead towards process optimization through AI and problem-specific models.
“This is much more helpful than asking ChatGPT a question. You can use AI to identify where bureaucratic processes can be optimized,” saving time and resources for everyone, the computer scientist said. This includes combatting issues like staff shortages as Germany experiences demographic decline. The aging population provides fewer and fewer people who will want to work in public administration, particularly when working in industry can come with higher rewards.
He added that this process was in some ways inevitable. “I predict that in 10 years, AI technology will be used for important tasks in all public authorities and at all levels of government.”
Some of these types of AI models are already in use in Germany, also at the local and municipal levels. The city of Cologne has a pilot project to use AI for better city planning and traffic management, while Munich is using it to create better garbage collection routines. In Heidelberg, it is being trialed for the early detection of diseases.
However, legal scholars and ethicists are still questioning whether the rush to implement AI at all levels of government is being done with enough oversight and forethought to the consequences of , especially when it comes to its use in healthcare, the judicial system and immigration procedures.
“AI can serve as a tool, for example, in preliminary examinations or standard cases,” wrote law student Jan Christian Swoboda on legal news site LTO. “However, human review is essential, especially in atypical cases or cases highly relevant to fundamental rights.”
Open AI/SAP deal raises questions over German tech independence
Germany has strict data protection regulations that apply, in particular, to implementing AI systems in the public sector. Personal data may not be stored longer than necessary and must be protected against unauthorized access, loss or destruction. Data subjects have the right to request information about what data has been collected about them and for what purpose, and they have the right to have unlawfully collected data deleted or corrected.
Hoos warns that despite promises of “sovereignty” via a Germany-based cloud, a deal announced last week between ChatGPT creators OpenAI and Germany’s software giant, is creating a “dependency on actors we cannot fully trust” and does not provide transparency as to whether OpenAI will have access to the data generated by SAP’s models or not.
“This involves public money that should not be spent on questionable American technology, but should rather be spent on building up EU and German AI capabilities,” he argued.
If, as Minister Bär’s “High-Tech Agenda” declares, Germany seeks to become a leader in AI, integrating it into homes, healthcare, and even the administration of justice, experts caution that independence is crucial.
Edited by Rina Goldenberg
While you’re here: Every Tuesday, DW editors round up what is happening in German politics and society. You can sign up here for the weekly email newsletter, Berlin Briefing.
The post How the German government is going all in on AI appeared first on Deutsche Welle.
