At least since ChatGPT, the term “generative AI” has been used more and more. But what exactly is generative AI and where is it used?
• Generative AI is a collective term for an AI that can generate something (e.g. a text, an image or a video)
• Generative AI is already being used in various areas
• Despite the many possibilities, there are also some risks and dangers associated with generative AI
What is Generative AI?
Generative AI refers to AI-based systems that are able to generate a variety of results in an apparently professional and creative way, as the Gabler Wirtschaftslexikon explains. These results include images, videos, audio, text, code, 3D models, and simulations, among others. Basically, any AI that generates something itself is a generative AI. So-called machine learning is used in generative AI. This is machine learning, which includes different forms of self-learning in artificial intelligence and robotics systems. In the case of generative AI, it is mostly about “deep learning”, whereby various data sources and training methods are included in the “learning process”. As the Industry of Things website explains, generative AI is all about improving current innovative workflows together with humans.
Generative AI can be used in these areas
At least since the AI boom, there has been a large number of generative AIs that are used in various areas. This includes not only large generative language models such as the voice bot ChatGPT, which attracted the most attention in the sector at the beginning of the AI hype, but also AI image generators, music generators, video generators and voice generators. Since the possible applications of generative AI are so diverse, they are already being used intensively in various industries. Typical uses include, for example, the creation of texts (e.g. web content, news texts, product descriptions or marketing texts), use as an intelligent chatbot for customer support, the creation of photo-realistic images (e.g. for advertising purposes) and support for programming new applications and creating program codes, to name just a few. Generative AI has even found its way into medicine. An example of the use of generative AI is the early detection of malignant tumors. The Generative Adversarial Networks (GAN) model uses different angles of an X-ray to calculate a visualization of the potential tumor volume. And generative AI is also already being used in art. So-called AI art generators can turn text into art or combine images into a new photo. And given its unique ability to generate synthetic data to train its own models, generative AI is considered one of the most promising advances in the artificial intelligence world.
A critical perspective on generative AI
In addition to technical progress, however, one must also deal with the ethical issues surrounding generative AI and look at the new technology from a critical perspective. Because despite the impressive possibilities that generative AI brings with it, it also harbors some dangers and risks. First of all, according to Bigdata insiders, it should be viewed critically that the output of generative AI can also be faulty. The AI tends to invent or “hallucinate” facts due to a lack of knowledge. In addition, content generated by the AI could be influenced by bias contained in the training material. There are also legal questions, for example authorship, liability and data protection, as well as the use of copyrighted material for training.
Editorial office finanzen.net