It’s not quite that simple. Crowdsourcing has many of the drawbacks that AI has too.

While it can have a higher reliability in detecting nonsensical inputs or inputs that it’s simply unfit in processing, that comes at an intrinsic cost in scalability. Some tasks can’t be effectively crowdsourced for, either because of volume or urgency.

Machine Learning systems learn to approximate decision making and thus can attempt at learning from crowdsourcing efforts. It is notable though that depending on the use case, model and training method, machine learning algorithms can potentially be better than the data it was trained on. Or much worse, it’s very fickle.

It is definitely still the case that crowdsourcing is a really important tool and oftentimes machine learning relies on it’s efforts. And it naturally can solve tasks that we don’t have a viable automated approach for.