randomly sort a list with bias

I have a list as follows:

i = [
        {'id': '1', 'P': 0.5},
        {'id': '2', 'P': 0.4},
        {'id': '3', 'P': 0.8},
        ...
        {'id': 'x', 'P': P(x)}
    ]

When I do the following:

i.sort(key=lambda x: x['P'], reverse=True)

The list gets sorted based on P where the element with the largest P is in front. but what if I want to make it seem randomized so that even a element with a small P value (with very small probability) can be the first item in the list?
Is it possible to implement that using the sort() function or do I have to write my own?