Reference guide · no prior knowledge
At my events I get lots of questions. Here are the most common ones.
Topic 1
AI can produce new combinations: texts, images, and ideas that did not exist before. But while you are talking with it, it does not learn anything new. It is trained beforehand, in an elaborate process that companies control.
Read answer →Yes, in two ways. For training, large parts of the public internet were collected. And many AI services now also search the web live before they answer.
Read answer →Yes, at its core that is exactly what a language model does. It calculates, piece by piece, which word most likely fits next. Whole conversations emerge from this, without the AI understanding what it is talking about, the way a human does.
Read answer →Yes. Books are among the most valuable training data, because they contain long, carefully built texts. In some cases, copyrighted works were used without permission, which has led to lawsuits.
Read answer →You only know once you check. AI phrases false statements just as convincingly as true ones. Checking means: look at the sources, search for a second independent source for important facts, and look especially closely at numbers.
Read answer →Topic 2
No. Today's AI has no goals of its own and no drive of its own. It becomes active when a person or a program starts it, and it only acts with the tools and rights that it is given.
Read answer →Yes, and you often do not notice. Many chat services automatically route your question to a fast model or to a more thorough one. AI agents also decide for themselves which tool to use for a given step.
Read answer →AI is already on Mars, though inside robots and not as a being of its own. Elon Musk had announced that humanoid robots would be sent to Mars on a Starship rocket at the end of 2026. In early 2026, SpaceX postponed this plan by years in favor of the Moon.
Read answer →Topic 3
Yes. Systems with autonomous functions are already in use, for example drones that steer themselves toward a target in the final stage of flight. There is no international ban so far. States have been negotiating about this at the UN for years.
Read answer →That depends on the service and your settings. With many free offerings, your input can be stored and used to train future models. With business versions, this is usually excluded by contract.
Read answer →Not the AI. Liability can fall on the manufacturer, the provider, or the person and company using the AI. Who is responsible for what depends on the individual case and is currently being newly regulated by law.
Read answer →Not without question anymore. AI now creates images, voices, and videos that are barely distinguishable from real ones. In the future, trust will come less from the image itself and more from the source that distributes it.
Read answer →Topic 4
Based on what we know today, no AI has consciousness. Whether that is possible in principle is scientifically open, because we do not even know exactly how consciousness arises in humans.
Read answer →It is more likely to change your job than to replace it. Individual tasks disappear, new ones get added. Activities that are mainly routine screen work are under the most pressure.
Read answer →The risk is real, but it is not fate. If you regularly hand off mental work to AI, you practice it less. What matters is whether you use AI as a shortcut or as a sparring partner.
Read answer →Send it to frag [at] rogerbasler [.] ch. Selected questions will be answered and added here.