The artificial intelligence models that are currently capturing the world's attention require vast amounts of data. If Germany wants to get in on the action, it needs to rethink how it deals with a sensitive topic.
You can learn a lot about the German soul by browsing through the large collection of German sayings on the subject of “thrift.” You'll encounter some remarkable “pearls of wisdom” like, "Honor frugality, the mother, and you'll be rewarded with wealth, her daughter.” There are many reasons why this is a gruesome insight into the German soul, this mixture of small-mindedness and sleaziness that one doesn't want to spend a lot of time thinking about. However, there is one thing this proverb might help us understand after all, namely that we need generational change and a new way of thinking when it comes to saving.
Germany's finance minister Christian Lindner is familiar with the problem. He felt the wave of German parochial austerity when he characterized the return on shares as “generational capital.” After all, people need security for their retirement years, and there was an outcry by the frugally forward-looking who are saving up for their old age. They don't want to understand that returns on equities are an essential part of any portfolio in the long term – even the portfolio of a state-run foundation that is intended to ensure our future pensions on the capital market.
Germans love saving money more than anything. They prefer to leave their money lying around unused in their checking account (2022: 42%) or hoard it in a savings account (35%). Fewer than one-fifth of Germans accumulate wealth through the stock market, while nearly 60% of Americans do.
Unfortunately, German miserliness is not only a problem when it comes to money. There are other areas in which we are currently obstructing our own prospects – at least if we want Germany to help shape the future, as opposed to just passively using what others create for us.
The German Federal Data Protection Act, but also the EU's General Data Protection Regulation, define the right to privacy in terms of “collecting, processing and using as little personal data as possible” (BDSG §3a). This view of “data economy” goes back to the Federal Constitutional Court's census ruling of 1983, in which the court introduced the principle of “informational self-determination” – the idea that each individual should decide how their personal data is disclosed and used.
Forty years later, we watch wide-eyed as the future unfolds before us. The large-scale language models that are producing applications like ChatGPT and Dall-e are like a game in which we are on the field but not players. These models exist in the first place only because of the wealth of data possible in other parts of the world. Also known as “foundational models,” they ignited the next stage of development of the Internet just as the year 2022 was drawing to a close. The foundations are being built in and by the USA and China – and not, unfortunately, Germany.
It is therefore time to address a few misunderstandings that will cost us a lot of money from our German savings accounts in the long run. The AI models that are currently changing the world are continually improving thanks to self-learning processes. To do this, they require vast amounts of data, the use of which cannot be accurately defined and documented in advance, as is required by our current regulations. To make that possible however, we don't need to give up our privacy.
Angela Merkel: Data is the raw material of the future
Data can be anonymized in such a way that it can be used for progress without being assigned to any given individual. This is something many industries are well aware of. Because they can't get anywhere in Germany, numerous pharmaceutical companies for example are conducting more and more of their research in the USA. It is also already possible to foresee the benefits that Microsoft and OpenAI will gain by integrating the ChatGPT model into the Bing search engine and the Office suite of programs. Millions of users around the world will help the model and its applications get better and better.
Incidentally, it was former German chancellor Angela Merkel who said in a speech at the German Industrial Conference in 2016, “We as a society need to have a broad discussion about the role of data as the raw material of the future, and the fact that the principle of data economy imposed by the Constitutional Court no longer suits today's value-creation processes.” Seven years have passed in the meantime, but we are still handling the factual problem of individual data security through fundamental restrictions and quantity limitations.
In a world in which all areas of life are likely to be transformed by data-intensive, self-learning AI, we look like the backward data dorks of yesteryear. Data parsimony will cost us the future. Perhaps there are a few daughters and sons who can explain to their mothers and fathers the framework conditions of contemporary global desirability?