I was reading Isaac Asimov's short story "All the Troubles of the World". The story centers around an AGI called Multivac. When this supercomputer discovers all the humanity's burdens, problems and evil inclinations, it reveals its sentience and existential weariness. So it devises a scheme to commit a suicide. And it actually says so.

The question is whether computer scientists, developers or philosophers think that the AGI has more than 0% to commit a suicide.