Question

So when a ML Model has high bias, getting more training data will help in improving the model

Answer 1

The given statement is true because we know that it is clear that more than one training data would help to lower the variance of any high variance model unless there would be less over fitting if a learning algorithm is open to more than one data samples.

Though high bias models would not benefit from the more than one training examples, there must not be any kind of consensus.