Düsseldorf, 4 May 2024

Long before OpenAI, ChatGPT, and all other generative A.I. applications, programs, or snippets hit the headlines of media worldwide, politicians and regulators in Europe and North America led a debate on how to define “trustworthy” and “human-centered” AI. They asked who should be the regulator, and what exactly should — sorry: can be regulated, and how this regulation could be enforced.

When I wrote my MBA thesis on AI regulation in Europe and Germany, four years ago, I looked at that question through the glasses of John Kingdon and his multiple‐streams theory which describes the world as three streams: problems, policies, and politics.

We are probably all aware that A.I is accompanied by a bunch of problems some of which we do not even know of. And we might raise doubts on whether we have got the right policies in place to rule A.I.. At the same time, it becomes clearer every day that if A.I. enters the political arena, politicians will (ab)use it for their own purposes whereas public opinion will probably be divided, polarized, or explosive.

Among the ideas I gathered during my research from sources of the European Parliament, one idea stroke my brain: Why don’t we create a legal person for Artificial intelligence?

You might think that’s absurd because A.I. should never get the same status as real humans. But I am not talking at all about a human-like natural entity – but compare the legal person of an “Artificial Intelligence Entity” (AIE) with the creation of legal persons for associations, and companies of all kind that triggered a broad range of laws governing these business entities.

We should be aware that we will need a package of laws governing the development, implementation, protection, and product liability for Artificial Intelligence Entities. Take a look at the progressive work in the European Parliament on creating a legal framework for A.I. But there is a missing block: the definition of a legal person. Sooner or later, law-makers will agree that this legal person is needed – and they will create a Law on Artificial Intelligence Entities.

I will dedicate a range of much more detailed articles, in the upcoming weeks and months, showing different cases of A.I. in practice hoping to prove that by creating a new legal person called an Artificial Intelligence Entity,  we will lead to much more effective and better regulatory and comprehensive solutions than the regulatory mish-mash we are facing at the moment, on different continents.

—

You can e-mail me at j.weidemann@comeconmedia.com