It's not like the biosphere has spent millions of years achieving a balance or that the balance is important.
Tell me, where is your evidence for there being a balance? I'm a geologist and I get paid (you know - cash, from businesses, for delivering useful product) for identifying the swings and surges in those reactions as the Earth's systems either fail to keep up with external changes to the conditions that determine that alleged "balance", or overshoot their adjustments.
You know how to balance a broom by resting it's handle in the palm of your hand, then jiggling your hand around to try to keep the broom head in the air. Yes? Done that? That's the sort of balance that nature has.
Or maybe that's a poor analogy. Try a double-pendulum : suspend one weight from an anchor, then a second weight from the first weight with either a rod (stiff, light, you remember the mechanics problems) or a cord (light, flexible) ; now agitate the anchor. Try to predict the motion of either mass in the pendulum.
Tectonic forces within the Earth (decreased heat flow as radioactive and construction heat decay) is one agitation to the system ; astronomical forces (the ~5%/ Gyr increase in heat flow from the Sun ; Milankovich orbital cycles) are other agitations to the system. There is no "balance, there never was, and until the heat death of the Solar System, there never will be. There are restoring forces, but there is no reason to think that they will be either large enough to stabilise the system, or small enough to not de-stabilise the system.
The best we can do is to try to reduce the forcings we are putting on Earth systems, or to apply the forcings we do control in directions that shift the current state in a direction that we want. We started to understand this (chemistry and thermodynamic systems) in the 19th century ; we started to understand the maths of deterministic chaotic systems in the late 19th and early 20th centuries (Poincare on orbits ; Mandlebrot in more general mathematics) ; we're still trying to get to grips with the wider implications.