Now look at population growth. The Roman empires population was largely stagnant between the 1st century AD to the 3rd. Africa's population has increased 10 fold in the same period.
What you say is partly true, partly not. Africans definitely die earlier.
But the current shape of the African population pyramid is partly determined by the baby boom in the last three generations. There were many fewer Africans around in 1963 (which is when people who are now aged 60 were born), so even if they all lived long (which they didn't), their share on the 2023 population would be much lower than in societies which barely reproduce themselves.
Has there truly been an African baby boom, or is it that the birth rate has remained stable-ish and the survival rate has skyrocketed?
I suspect it is a bit of this and a bit of that. It is not just child mortality, but mother mortality (during/after birth) which plummeted with better medicine, thus the same amount of women can bear more children.
You can't look at this without compensating for population growth and get a meaningful comparison. A raw population pyramid is only a reasonable-ish indicator of life expectancies when a population is growing at slow rate.
For a fast growing population you'll always see a significant narrowing by age no matter if people live long or not.
https://www.populationpyramid.net/africa/2022/
To compare to a time when slaves were near, surely we must add our poor regions, where people work in mines and factories for us, to our life expectancy. Otherwise we are as much comparing Modern Capitalism to feudalism, etc.