Artificial

Intelligence.

Artificial intelligence is affecting more and more business processes. At the same time, there are still many unanswered questions when dealing with AI. Our team of specialized IT and IP lawyers advises companies and entrepreneurs on the handling and use of AI.

Call now!

What is artificial intelligence

Artificial intelligence (AI) is, simply put, the simulation of human intelligence or competence by automated systems. This means that AI systems attempt to imitate skills such as learning, thinking or problem solving that normally only humans possess. Unlike traditional software, which consists of fixed rules and instructions, AI can learn from sample data and adapt its behavior to new situations. Conventional programs always deliver the same, predictable results when given the same input. AI systems, on the other hand, adapt their algorithms independently: They recognize patterns in the data and use them to derive rules for new data. As a result, AI applications are often more flexible and “intelligent” than conventional software.

How does AI work?

Modern AI is primarily based on machine learning (ML). Instead of programmers specifying all the decision rules, an AI model is trained with lots of sample data. It adjusts its internal parameters until it can solve the tasks set as well as possible. One common type is artificial neural networks – computational models that are modeled on the human brain. These networks consist of many virtual “neurons” that are connected to each other in layers. With enough data, they learn to recognize complex patterns, for example in images or spoken words.

Large language models (LLMs) are a special type of neural network. These are very large language models with billions of training parameters that are trained on huge amounts of text. An LLM such as OPenAI’s GPT (the model behind ChatGPT) learns statistical relationships between words and sentences. This enables it to generate probable continuations from a given sentence beginning. In this way, LLMs answer questions in natural language or write coherent texts that often sound surprisingly human-like.

Areas of application for AI

AI technologies are used in many areas today. Here are some important areas of application:

AI in companies

AI offers companies enormous opportunities for automation and increased efficiency. For example, it can automate repetitive processes and analyze large amounts of data in order to gain valuable insights. AI-supported chatbots are used in customer service, which can answer questions and solve problems around the clock. AI can also help in marketing, for example by providing personalized product recommendations or addressing target groups more precisely. Overall, companies can use AI to make decisions faster, reduce costs and develop new business models.

AI for creative applications

Artificial intelligence also opens up completely new possibilities in the creative field. Generative AI systems can create content independently – for example generative art, in which algorithms generate images or graphics that look as if they were created by humans. The same applies to music: AI can compose melodies or generate beats that serve as inspiration for artists. In design , AI tools support designers by suggesting layouts or color palettes. And in text generation , language models help authors by drafting texts or assisting with translation and proofreading.

What is important is that AI will not replace human creativity, but will serve as a tool for finding new ideas and relieving the burden of routine work.

AI in the legal sector (Legal Tech)

In the legal sector, AI is used to support legal activities under the heading of legal tech. For example, specialized AI programs can analyse documents and contracts by automatically checking contractual clauses for risks or irregularities. When analyzing documents, AI sorts and searches large volumes of legal texts (e.g. court rulings or legal texts) for relevant information in the shortest possible time.

Here too, AI can support attorneys and relieve them of routine tasks. This gives lawyers and legal professionals more time to focus on complex legal issues and advising their clients.

Generative AI: examples of text, image, video and music generators

A particularly rapid trend in recent years is generative AI – AI systems that can independently generate content such as texts, images, videos or music. Such systems learn from existing data (e.g. thousands of images or books) and can create new works from it. Here are some well-known providers and models of generative AI in various media sectors:

Text: ChatGPT (from OpenAI), Claude (Anthropic), Gemini (Google) and Le Chat (Mistral) are AI language models that understand human language and can formulate texts themselves. They are used, for example, for text generation, summaries, translations or as virtual assistants.
Images: DALL-E (OpenAI), Midjourney, Stable Diffusion, Flux and Adobe Firefly are so-called text-to-image generators. They generate artificial images and graphics from written descriptions. This means that even non-experts can create impressive illustrations or works of art with just a few words.
Videos: Runway, Synthesia, Adobe Firefly and Dream Machine offer AI tools for video generation or editing. They can be used to produce short video clips, animations or virtual presenters (avatars), for example, without the need for a real film crew.
Music: Suno and AIVA are examples of AI that composes music. They can generate their own pieces of music based on learned musical styles, which can be used for background music or as a source of inspiration for musicians.

The problem of hallucinating AI

Hallucination describes the phenomenon of AI models generating content that appears plausible but does not correspond to reality. In particular, large language models (LLMs) such as ChatGPT can present invented facts, false sources or even completely new concepts that do not exist. This problem arises because AI does not “know” what is true or false – it only recognizes statistical patterns and probabilities in texts. Critical review and human control of AI-generated content is therefore essential, especially in sensitive areas such as law, medicine and science. It is therefore essential to check content for accuracy.

Ethical and legal issues

The spread of AI technologies brings with it a number of ethical and legal issues as well as opportunities. On the positive side, AI helps to speed up many processes, reduce errors and find new solutions to social problems – be it in medicine, environmental protection or relieving people of monotonous tasks. At the same time, however, the risks of AI must also be taken into account:

Data protection: AI systems often require large amounts of data. There is a risk that personal data could be misused or incorporated into the training data without the knowledge of the data subjects. It must therefore be ensured that the applicable data protection laws are complied with and that sensitive information remains protected.
Transparency: Many AI models – especially neural networks – are so complex that their decision-making processes are difficult for humans to understand (keyword: black box). This lack of transparency can be problematic if, for example, an algorithm rejects a loan application or makes a medical diagnosis without it being clear why. It is therefore required that AI systems provide explainable results or at least disclose where AI is used.
Bias and discrimination: Another aspect is bias, i.e. the distortion of results. If an AI is trained with biased or discriminatory data, it can adopt these prejudices. For example, there have been cases in which algorithms discriminated against women in job applications because the training data came from the past, when fewer women were hired. Recognizing and preventing such biases is a major challenge if AI is to remain fair and equitable.
Responsibility: Ultimately, the question arises as to who is responsible for the decisions made by an AI. If a self-driving car causes an accident or an AI makes the wrong financial decisions, who is liable? Developers, operators and legislators need to clarify how to deal with such cases. The aim is to create mechanisms that prevent wrong decisions and clearly assign responsibilities in an emergency.

Despite these challenges, AI offers great opportunities for companies, but the technology must be used responsibly. Companies need appropriate guidelines and controls to ensure that AI can be used in a way that is beneficial without causing harm.

The EU AI Regulation

In order to strike a healthy balance between the opportunities and risks of AI, the European Union has adopted binding rules for AI, the so-called AI Regulation. This EU regulation is intended to create a uniform set of rules on how AI systems may be developed and used in Europe. The AI Regulation takes a risk-based approach: AI applications are divided into different risk categories depending on the potential risk – from low to high to unacceptable.
Applications with an unacceptable risk (e.g. manipulative social scoring systems or mass surveillance) are therefore prohibited in the EU. High-risk AI systems (e.g. in areas such as medicine, transport or justice) remain permitted, but must meet strict requirements. Providers of such systems must, for example, carry out risk assessments, ensure transparency and demonstrate compliance with security and data protection standards. Less risky AI applications, on the other hand, should primarily meet transparency obligations – such as labeling when users interact with an AI system or when content (such as texts or images) has been generated by AI.

The EU’s AI Regulation is the world’s first major legal framework for artificial intelligence. The regulations will come into force gradually: initially for prohibited applications and training obligations, and later for high risks and transparency obligations. For companies, this means that they will have to review their AI systems and, if necessary, adapt them to the new requirements in order to remain legally compliant. In the long term, the AI Act is intended to ensure that AI is used in Europe in a trustworthy manner and in accordance with European values. It is therefore an important step towards exploiting the opportunities of artificial intelligence and limiting its risks.

We advise and support companies on all legal issues in connection with artificial intelligence (AI), both providers and developers of such systems as well as customers who use such systems or those whose laws have been infringed by artificial intelligence (AI).