The end goal is proprietary lock-in, AKA The Microsoft Strategy. The reality of auto-ML is that you don't need garbage data to learn garbage. All you need is

The end goal is proprietary lock-in, AKA The Microsoft Strategy. The reality of auto-ML is that you don't need garbage data to learn garbage. All you need is a "developer" who has no idea how dirty or unstable their data is, how accurate/precise the results should be, or where bias and edge cases manifest before GIGO becomes your project's mantra. The claims of auto-ML and auto-programming are two sides of the same coin. They're both silicon snake oil.