mtDNA can be used to track lineages. It's the dating of when branches occurred that is off because it assumes all differences are due to mutations and not paternal contributions.
With normal DNA the crossover add some changes and in particular it mix the changes in one branch of the tree with the changes in the other branch of the tree, so it's not longer a tree.
So my guess is that it will not change the dates too much, specially since he dates are calibrated by historical events. With made up numbers, something like "we know that this population reach here 10000 years ago and has 37 changes and that population reach there 20000 years ago and has 69 changes. This other population has 53 changes so after some complicated math, we estimate that they split 15000-1600 years ago.
I guess that the size of the population at the bottlenecks will be slightly reduced, because IIUC they are not calibrated using historical data but using some mathematical models about how the size of the population affect the diversity. I don't expect a big change nevertheless.
The rate of mutation is already not a constant, and so such dating is always approximate, there is a margin of error.
If you wish to demonstrate that it is now "completely off" and "invalid", try harder. I call shenanigans.
This is a drastic and totally implausible statement that is not supported by any serious geneticist, nor by the Wikipedia article you've linked to.
mtDNA does not have to come exclusively, 100%, no-exceptions-ever from the mother, in order for population research which uses to be valid! 99% is quite sufficient. It is at least 99.9% accurate: as the Wikipedia article helpfully points out, there are 1000x as many mitochondria in an egg as in a sperm.