Germany looks south. Access to ChatGPT was blocked last week by Italy's data protection regulator, so it can't be used in that country (although there are ways around the block). German authorities have not yet begun to turn off the tap, but have asked their Italian counterparts for information as debate rages in the country over whether such a move is appropriate.
The only clue to that possibility was given by German data protection commissioner Ulrich Kelber, who admitted his country could follow Italy. “In principle – he qualified – a similar procedure is also possible in Germany”. At the moment, everything is in the study and information gathering phase. In addition, the federal states also have competences in the matter.
The federal government seems little predisposed to start blocking the use of an AI in the country. The Federal Ministry for Digital Affairs and Transport stated: "We don't need to ban AI applications, what we need is to find ways to guarantee values like democracy and transparency." The department head of that ministry pointed out that the future legal framework of the European Union for this technology will make Europe a "world pioneer of reliable AI".
Another German federal ministry, the Interior Ministry, claimed to be "closely following" the information published by the media about the Italian blockade in application of the General Data Protection Regulation (GDPR) of the European Union.
The EU is also in this sleeplessness, which is processing a bill on artificial intelligence that should be concluded throughout this year. Emergencies are much more pressing if possible in a matter of such rapid evolution and diffusion among the population. It's only been three weeks since GPT-4, the latest OpenAI AI language model, was released, and there are already thousands of different applications for a multitude of fields.
Brussels, at its pace, is aware of the magnitude of the risks it faces. The European Commissioner for the Internal Market and Services, Thierry Breton, explained to the French station on Monday that he intends to include in the future European law on AI the requirement that any product generated with this technology bear an identifying mark.
“It is an extremely important issue – warned Breton – that is why I am promoting the Artificial Intelligence law, which is currently being debated in the European Parliament and which should be voted on this month”. “Everything that the AI generates will have to be marked as made by an artificial intelligence,” she concluded.
The European commissioner hopes that the text can be voted on at the end of this month in the European Parliament. The implementation of community regulations, however, meets generous deadlines, so it is possible that this new legal framework for AI will not be in force within this year or the next.
At the moment, in Italy, OpenAI has a little less than three weeks since its blocking last week to demonstrate to the data protection authority that European legislation has not been violated. The CEO of the company, Sam Altman, assured that he believed that his company complied with the European regulation and expressed confidence that they will be able to prove it.
At the moment, the company is returning the amount paid by paying subscribers to ChatGPT. The paradox is that Microsoft's Bing search engine, which uses GPT-4, can be used normally in Italy.
