Quantum computers have the potential to revolutionize quantum chemistry because they can, in principle, compute accurate electronic structure models very efficiently. At the moment, however, we are technologically limited to so-called ‘noisy intermediate-scaling quantum' (NISQ) devices, which suffer from calculation errors (noise) because they are very sensitive to their environment. Despite these noise issues, accurate molecular energies can be obtained on NISQ devices, but unfortunately only when the underlying wave function is variationally forced to overcorrect for the computational errors made by the quantum computer. In this project, we will show that relying on such error corrections negatively impacts the quality of the wave function and causes unreliable performance for general chemical applications. By analyzing how quantum computing wave functions describe the atoms in the molecule, we will uncover their chemical content and will determine if they erroneously predict molecules to dissociate into fractionally charged atoms. Moreover, we will demonstrate that atoms as conceptual open quantum systems offer the ideal tools to solve quantum computational errors in an innovative way without compromising on quality. In this way, we intend to create the first 'killer quantum computing app' that allows researchers to fully apply the power of quantum computing to their research without sacrificing the chemical concepts they hold dear.