Only if you ignore the unsolved and persistent MAJOR, high risk,problem:  It doesn't actually work.  No AI model is accurate enough to avoid disaster: there will be countless stories of AI telling your boss something you didn't want said, or making purchases you didn't want, telling you to add glue to pasta sauce. These problems aren't going away, and the only way to stop such risks is to make AI a less useful assistant, which is what a standard computer is.