https://en.wikipedia.org/wiki/Thermoelectric_effect
Instead of that, probably a more feasible way to make use of heat is just as heat, by some way of circulating it into buildings in the winter. One interesting angle of looking at this would be that it's really an artificially created geothermal source.
Your AC is working hard at pushing all the heat outdoors and making a pocket of cool inside.
Air conditioning as we understand and use it today is a gold-plated band-aid over bad design. It's so entrenched you almost can't get a house built any other way.
The basic idea is that the thermal time constant of a uniform sphere (or other shape of fixed aspect) is proportional to the square of its radius. So, for a sufficiently large sphere, the time constant can become very long, >> 1 year. Rock is a good choice because it's very cheap.
The other thing one can do is store cold. This is useful not only for cooling, but for increasing the efficiency of storage of work, since one can generate heat and cold via a heat pump, then reverse that cycle to recover the work. An efficiency of 65% or so may be reasonably achievable with the high temperature in the temperature range of ordinary steel.
Shouldn't the thermal mass go like volume (radius cubed) while the power conducted away go like surface area (radius squared) so the time to cool should go like mass/power ~ volume/area, so linearly in radius?
https://en.wikipedia.org/wiki/Seasonal_thermal_energy_storag...
I was especially surprised to learn that serious attempts at storing heat in soil and rock have been made
https://en.wikipedia.org/wiki/Seasonal_thermal_energy_storag... https://en.wikipedia.org/wiki/Drake_Landing_Solar_Community
which in the winter is able recover >40% of the energy stored in the ground through boreholes in the summer ("BTES Efficiency")
https://www.dlsc.ca/reports/swc2017-0033-Mesquita.pdf