The problem is that Earth's rotation isn't consistently faster. Some years leap seconds need to be added, some years they need to be removed. Would be far better to leave them alone, let them average out, and as the GP said let the people who care about this add the offset they need.
> Some years leap seconds need to be added, some years they need to be removed.
Is that true? Per Wikipedia:
> Since [1972], 27 leap seconds have been added to UTC, with the most recent occurring on December 31, 2016. All have so far been positive leap seconds, adding a second to a UTC day; while a negative leap second is theoretically possible, it has not yet occurred.
Either way, it's due in part to Earth's rotation slowing down, so the average drift would still be non-zero.
We've not had to apply negative leap seconds yet since leap seconds were introduced in 1972, but that wasn't the point.
The time period of the Earth fluctuates a lot [0] and actually in 2020 it was less than 24 hours, but not a large enough change to warrant a negative leap second. If you go back to the 1940s, we would had needed negative leap seconds if we had leap seconds at all then, and going back 150 years we would have needed multiple negative leap seconds every year for several consecutive years.
What we can say is that on average, it is close enough to 24 hours and the average over hundreds of years is even closer to 24 hours that it's not worth adding these extra seconds as you'd then need to remove them again later on.
[0] https://c.tadst.com/gfx/900x506/graphlength-of-day.png from https://www.timeanddate.com/time/negative-leap-second.html