Artificial intelligence can acquire scientific knowledge

This is nothing like what we might remember from school lessons. There one can know Kepler’s third law of planetary motion mostly in this form: “The squares of the orbital periods are in the same ratio as the cubes of the semi-major axes.” square of the orbital period (s) and the cube of the planet’s distance from the sun (any) can be seen in the formula.

However, the AI ​​equation has nothing to do with Kepler’s third law as it is usually known. In its exact form, one must of course also take into account the mass of the sun and planets there, as well as the gravitational constant. Mass, the gravitational constant, and other numerical factors can all be summed up in Kepler’s formula as 0.1319 (and then you have to keep in mind that the AI ​​didn’t calculate it in the usual SI units of meters, kilograms, and seconds for technical reasons). But this equation is far from a real “law of nature” or knowledge that comes close to that of Johannes Kepler.

Artificial intelligence should not be equated with Kepler or Newton

It starts with the basic requirements: the AI ​​is not only given data on orbital masses, distances and times, but also various formulas, including Newton’s law of gravity. Unlike Johannes Kepler, artificial intelligence “knew” how celestial bodies attract each other. If Kepler had this knowledge, he could have saved working with his three laws. Because it is just another formulation of what Isaac Newton discovered later (more precisely). The formula developed by artificial intelligence for the relationship between orbital time and distance is also missing some physical content. Precisely because factors such as the gravitational constant or the masses involved are not explicitly listed, it is almost impossible to derive further insights from them.