Richter und Roboter by ARIC x Fieldfisher: Wer in Deutschland

Judges and robots: data protection as a wonder weapon

Wie beeinflusst ChatGPT die Jurawelt? Werden Urteile bald nur noch von Robotern gesprochen? Und was sagt eigentlich die Rechtssprechung zu den neuesten Entwicklungen der generativen KI? In der neuen Kolumne Richter und Roboter von ARIC und Fieldfisher beleuchten Dennis Hillemann und Stephan Zimprich im Wechsel einmal monatlich Themen aus KI und Recht. Im Februar geht es um Datenschutz und die Einschränkungen, die er mit sich bringt.

Phew, just got lucky again. Progress is not unstoppable, the legal profession is saved. The Federal States Task Force on Artificial Intelligence has pulled the super joker and put the data protection officers of the federal states on OpenAI. But is that really a good thing?

A few weeks ago I tried out a tool called CoCounsel, and for a few days afterwards I was firmly convinced that I had seen into the abyss. CoCounsel is based on GPT4 and can research, summarize, analyze contracts and output the results in nicely formatted memos and tables. As a test, I had two memos created on topics I am currently working on. The results: shockingly good. If it had taken an associate three or four hours, I would have had nothing to complain about. It took CoCounsel three minutes. And before someone comes around the corner again and says well, it’s in English, and the tool doesn’t know German law either – yes, that’s true, but it’s clear that something like this is also coming for the German and European market, and in months rather than years.

Lawyers are conditioned to think in worst-case scenarios. For their own business model, in a nutshell: Lawyers in most larger law firms sell time. Their own time, partners then also the time of their associates. If a client gets the same answer from a software in a few minutes in the future for which they have to buy several hours from an associate today, then it’s clear that this will not continue. I then dug out a few old plans. Open a bicycle repair shop, become an artist, that sort of thing.

 


About the author

Stephan Zimprich
Stephan Zimprich
specializes in cases with a technology background and advises clients mainly from the digital sector in the areas of data protection, competition law, media law and IT law.

 

 


 

But you can rely on Germany after all. If digitalization threatens to advance, then bang, miracle weapon data protection. For me, this quickly leads to a pulse, because it has so far stopped every attempt at innovation and reliably “protected” authorities, schools, hospitals, universities and courts from dangerous digital improvement. Thesis, only slightly exaggerated: If you want to become something in Germany, you first have to get past the state data protection commissioners. “The widespread roll-out of AI flying blind – without a legal basis – is the main problem,” says Dieter Kugelmann, head of the Federal States Task Force for Artificial Intelligence and State Commissioner for Data Protection and Freedom of Information in Rhineland-Palatinate since 2015, announcing a coordinated approach to OpenAI by the state data protection authorities. It’s good that you need a legal basis for innovation in Germany. Entrepreneurial freedom, that would be even better, where are we going, and everything was better in the 1970s anyway, when there were hardly any computers and hardly any data, a dream. Germany is putting up the data protection fence. Everything is the same as always, whether it’s Schrems 2, Office 365 or teaching via video conferencing during the pandemic – “but data protection!”, nothing simpler than prevention, is after all a ban with a reservation of permission

Is that all right? Difficult. Language models are powerful, of course. And the speed of development is enormous. Slow it down a bit, if you like. Nevertheless, the plans for the bicycle repair shop are back in the drawer. AI makes lawyers more efficient, of course. That can also release energy. Can services that were too expensive “before AI” to really attract clients now perhaps be offered at attractive conditions? Are new offers creating new needs? This is a challenge. We have to be creative and we will try out new things. Some will fail, some will work. Doesn’t sound bad, I think. And I’m already annoyed with data protection again, Germany 1970, they’d probably like that.

 


Further episodes of the column: