I don’t believe large language models genuinely “think”—based on my experience, they spot patterns and replicate styles but never understand or reason as we do.
Neuroscientifically, human reasoning is both distributed and dynamic: it draws on an interplay between specialized regions like the prefrontal cortex, where we process logic, abstraction, and decision-making, and other networks involved in memory, emotion, and social intuition. What makes human reasoning unique is our ability to consciously create mental models, consider hypothetical scenarios, weigh causal relationships, and reflect on meaning—often synthesizing diverse inputs in a way machines cannot replicate yet.
Whenever I’ve used AI, whether to write emails or compose music, like in our Beethoven X project, the technology extends existing fragments and delivers plausible results, but these are always based on learned templates, not new ideas or deep insight. In the Beethoven X project, for example, the AI filled in missing pieces using musical logic drawn from Beethoven’s time. It even created unexpected continuations, but it cannot make any logical deductions about formal disposition or large scale developments, let alone reference something outside of the pure musical material it was trained on.
Why does this matter?
True creativity or reasoning means connecting unexpected ideas, questioning assumptions, and inventing something that never existed before. AI tools can support our routines and offer inspiration, but real understanding, intention, and invention remain human. My best results come when the machine helps me generate options, and I then apply my judgment—selecting, combining, or reimagining in ways no algorithm can anticipate.
The Interview was done by Michael Sowa for the Rotary Club Salzburg Altstadt in German. The video was originally in German, English version by @elevenlabsio If you want to watch the German original, you can find it here: https://youtu.be/kEKfFNKwQZ4
Oh, and in reality I was having a beer, not water, but that was changed by the photographer, not me, with another AI. I would never do that ;-)
#AI #Creativity #ArtificialIntelligence #Originality #FutureOfWork #TechEthics #MusicAndAI #DigitalLiteracy #AIReasoning #TechnologyAndSociety #HumanVsMachine #Rotary #RotaryInternational
Reasoning is beyond current AI system capabilities. Don't let the marketing blurb fool you.
LLMs cannot think. They cannot consciously create mental models, consider hypothetical scenarios, weigh causal relationships, and reflect on meaning. Talking about reasoning is only marketing blurb. Don't fall for it.