Artificial intelligence (AI) has never been more popular or talked about than it is now. AI tools, such as AI chatbots, have been widely adopted and are used to have conversations with or to gain knowledge about certain topics. Artificial intelligence is not new, and is used more than we realize, but its quick evolution and increased use cases bring with it legal implications for users and developers of AI alike.

What is AI?

Artificial intelligence is intelligence displayed by machines in a manner that is similar to human intelligence. As a result, an AI system is a system that can learn to recognize patterns, display information, produce creations (such as images, song lyrics or papers) and have conversations.

Traditionally, a distinction is made between strong AI and weak AI.

Strong AI is a form of AI that can learn and think exactly like humans and is not restricted to one task. Strong AI would be able to ‘think’ for itself and go beyond the boundaries set for it. In that respect, strong AI mimics human intelligence and can think and learn in a way similar to how humans think and learn. As of yet, there are no known examples of strong AI.

Weak AI, on the other hand, is created and trained to perform specific tasks. While it may sometimes not look or feel like it, all AI tools that are currently on the market are classified as weak AI. Even DALL-E, which can generate complex, creative images, or Chat GPT, which can produce long texts and can converse with users, are examples of weak AI, which are trained to do only one task (generate images or produce text).

The categorization as ‘weak’ AI is sometimes considered to be misleading, in that there are types of weak AI that can outperform humans and perform tasks in a quicker and more efficient way. For that reason, weak AI is sometimes more accurately referred to as narrow AI or specialized AI. For example, DALL-E can generate images quicker than any human can.

How does AI work?

AI works on the basis of data and algorithms. Datasets (for examples pictures) are fed into the AI system and by way of the algorithms, the AI system learns to recognize the patterns it is programmed for. In the example of pictures, tags will be applied to the pictures so that the AI system can learn from them and recognize the pictures (for example: pictures of blueberry muffins and the faces of Chihuahuas). Once it has been fed with datasets, new data will be introduced to see if the AI system can accurately apply what it has learned (for example: correctly distinguish between a blueberry muffin and the face of a Chihuahua). If AI system can accurately distinguish the patterns, it is ready to be used in the real-life situations it is created for.

What are the legal implications of AI?

It is impossible to summarize in one blogpost all of the legal implications of AI. That is why we will be delving deeper into the legal implications relating to various carefully selected and hot-topic legal issues in a series of blogposts that will be posted in the coming weeks. The following matters will be covered:

• AI and ethics
• AI and the proposal of the AI act
• AI and product liability
• AI and intellectual property rights
• AI and GDPR
• AI and consumer rights

In case you have any questions in the meanwhile, relating to the aforementioned or other topics, do not hesitate to reach out to us.

For additional reading on AI, we refer to our previous blogpost about the legal implications of AI Chatbots.