● news.bsdnet.hu

Understanding how generative AI works can help you recognize misleading content and make better decisions about what to trust. The following learning materials are from Lesson 15 of Detekto.hu, a free e-learning toolkit developed by the Hungarian Idea Foundation for students aged 13–18. This English-language module focuses on AI and misinformation. AI was already part of our lives long before we started generating funny images with Midjourney, creating realistic videos with Sora, or doing our homework with ChatGPT. Artificial intelligence refers to computer systems that are capable of performing processes similar to human thinking . For example, they can learn from experience without being programmed step by step, draw conclusions from data, and make decisions. Today, the biggest online platforms could not function without AI. For instance, eBay uses AI to recommend products, while TikTok and Netflix rely on AI to decide which videos or movies to suggest next. In transportation, it helps plan traffic or even controls self-driving cars, and in healthcare, it supports the detection of diseases. Chat programs that write homework for you. Image generators that create realistic pictures of almost anything — when people talk about artificial intelligence today, these eye-catching tools often come to mind first. These platforms are powered by so-called generative artificial intelligence . Generative AI models create something new: they write text, generate images, produce computer code, compose music, or create videos. The models behind these systems do not understand what they write or create. They try to predict, or more accurately calculate, based only on the data they were trained on, which words can appear in succession in a text, or which colors of pixels can be placed next to each other in an image or video. The “knowledge” of generative AI is based on massive amounts of text, images, audio, and video. The system looks for patterns in this data, establishes connections, and when you ask it something, it generates a response based on these patterns. For example, if you ask ChatGPT to write an essay about the situation of bees in Germany, it will start generating a response within seconds. But it does not actually understand what it is writing. It simply calculates, based on previously seen information, which word or sentence is most likely to come next—until the text is finished. Image generators work in a similar way. They don’t really understand what a spaceship is. But by analyzing countless photos, videos, scientific texts, and news articles, they have learned what shapes and colors usually appear in images of spaceships. So when you ask for an image of yourself as an astronaut sitting in a spaceship, the system only needs a photo of you, and then generates the image. Chat programs that write homework for you, image generators that create realistic pictures of almost anything – when people talk about artificial intelligence today, these eye-catching tools often come to mind first. The platforms are powered by so-called generative artificial intelligence. The goal of generative AI is not to tell the truth, but to give you an answer that seems believable and useful. It will respond even if it does not actually have enough information to answer the specific question. Because of this, it can sometimes “ hallucinate ”, meaning it provides information that is completely made up but sounds real. It often presents these false details very confidently, so you may only notice the mistake if you know the topic well – or if you check reliable sources. You also cannot be sure how reliable the data was that the AI learned from. It may include errors, biases, or inaccuracies. In such cases, the generated answer may also be biased or incorrect . AI-based models can be useful in many ways. However, the same tools that can help with learning or creativity can also be misused. Generative AI makes it possible to create misleading content quickly, cheaply , and even in a personalized way. Text generators like ChatGPT can be used to create fake news that supports a specific viewpoint. Image generators like Midjourney can create realistic images of events that never happened, sometimes putting real people in a bad light. As a result, people can become confused, as they find it increasingly harder to distinguish between real and fake information. Trust in news decreases, while fear and anger in society can grow. There is no single clear sign, but there are warning signals : Technology is developing fast, and the generated pictures and videos are becoming more realistic every day. That’s why you need to use your common sense to recognize AI-generated content. AI-based models can be useful in many ways. However, the same tools that can help with learning or creativity can also be misused. If something seems suspicious: Generative AI is not good or bad on its own. It depends on how it is used. The better you understand how these contents are created, the less likely you are to be mislead.