? The aim of this lecture and of the next lecture is to provide you with the framework for the units that follow later on in the course. I'm going to do this by focusing first on the legal significance of artificially intelligent software, so artificially intelligent computer programs in the next lecture, we will then consider the sometimes overlooked legal significance of the hardware dimension of AI, the computer infrastructure itself, on which artificially intelligent software runs. With respect to the artificially intelligent software itself, this lecture will ask and try to answer two questions. First, what is an artificially intelligent program? And, second, how does law relate to artificially intelligent programs? Let me begin with the first question. What is an artificially intelligent program? Unfortunately, this is quite a complicated question, and there are many different ways to answered. So it might be helpful to break this question down just a little bit further and to look separately at the meaning of the terms intelligence, artificial and program. The term intelligence has a long and complicated history. The word itself derives from Latin inter, which means between, and Legere, which means to choose or literally to read. So one could say that being intelligent means literally to be able to draw distinctions between different things to understand or to comprehend oneself and the world around us. The term artificial intelligence assumes that this human ability to understand, to comprehend, to sort the important from the unimportant can be replicated by constructing computer programs that are as good or sometimes even better than humans at understanding. Sorting, comprehending a given state of affairs. Just how powerful such artificially intelligent programs are differs quite a lot. Some programs can only perform very basic tasks like adding numbers, and these programs might not really deserve to be called intelligent at all. But other programs can perform very complicated tasks, such as playing chess or simulating complex medical treatments. At this stage, I will only mention three examples of such AI programs, but you will encounter many more in the course of the smoke. The first example. Are artificially intelligent programs that can be used for the purposes of facial recognition when searching for criminals in public places, for example, you will hear more about this in Ulrika Andersson's lecture the second example, artificially intelligent programs that can be used to process and analyze large amounts of written or spoken language. Such programs are helpful for understanding how modeling natural language works. Jeffrey Attic will focus on aspect related to these issues in his lectures of this movie. The third example. Are artificially intelligent programs that can be used to optimize the pricing of products or services. One Sales online, for example, by determining the best time in the manner of a transaction, Julian Nowak will address such situations in his lecture. As I mentioned, these are just three examples, and you will encounter many more applications of AI. In the course of this mook, the diverse applications of AI that we will present in this course often enhance our quality of life and help humans to make sense of the world we live in. At the same time, however, the application of AI, it's not always unproblematic. And with this I am turning to the second question of this lecture. How does law relate to AI? At present, we are already surrounded by many applications of AI in our computers, phones or in public places, and in many ways, our legal systems are well equipped to deal with these innovations. Just as our legal systems have dealt with other innovations like the Internet or the telephone in the past, however, that are at least three aspect of AI that are of concerned. From a legal point of view, the first aspect concerns the fact that there's a risk that the preferences, the interests the biases of. Those who write artificially intelligent programs are reflected in the programs themselves. To some extent, this is unavoidable. Our convictions and our interests always color what it is that we do. However, this circumstance is particularly problematic in the context of AI, since AI tools can be so very powerful. For example, It's one thing if one police officer treats suspects differently based on ethnicity, it is quite another thing. If the police generally were to use an artificially intelligent software that would systematically trade suspects differently based on ethnicity, the concerns in this particular area are very real, since many facial recognition programs, for example, failed to function properly when applied to non white ethnicities. From a legal point of view, this means that it is important to be aware of the discriminatory potential of AI applications. AI is not necessarily more or less biased than its creators. But due to the fact that AI programs can be applied easily on a much larger scale, AI amplifies the biases of its creators. A second aspect concerns the difficulty of supervising AI programs. When a human being makes a mistake, it is easy to say that the human being should be held accountable for that mistake. When a machine makes a mistake, for example, when a car breaks down, one can often argue that the person who built the machine should be held accountable. However, with more advanced AI programs, this is not as easy. And this is so because it is not always possible to predict how an artificially intelligent program operates in detail. And it is also not always possible to explain reach, respectively, in hindsight, how and why an artificially intelligent program came to the conclusion it did in the health care context. For example, certain algorithms are able to predict the best way of treating a patient, but it is very difficult to understand, just on which basis those algorithms make these recommendations. And that is why, from a legal point of view, one must think very carefully about how the running of AI program is supervised and who should be responsible when things do not go as planned. The third and final aspect that I would like to problematies from a legal point of view concerns the why question why and for which aims is a particular, artificially intelligent program used. AI can be used to control people, for example, when tracking how humans behave. And it can also be used to enable people, for example, when enhancing people's mobility with the help of autonomous cast. In part, this. Why question the question Why we use AI is a political and philosophical question, but it is also a legal question, since it determines if AI programs should just be treated like any other invention, like a photo camera, for example, or if it should be treated in a much more cautious manner to reflect the high stakes that are involved when using AI. As so often in legal contexts, there are no clear answers to these questions, But one core aim of this course is to also enable you to make up your own mind about how you think AI should be regulated so now we talked a little bit about the definition of AI and how law relates to it. And understanding these aspect is one important precondition for forming your own opinion on AI. But another and often overlooked aspect of that question concerns the actual hardware on which AI runs, and this is what we will look at next.