We could do it, but it's probably not an optimal solution for a couple of reasons. For one thing, as others have noted, deserts are actual functioning ecosystems: even the Sahara has native animal life, and doing a massive land use geoengineering project would disrupt those ecosystems, with somewhat unpredictable results. Even setting that problem aside, though, there are some substantial issues. All of the world's terrestrial ecosystems put together remove about 3 gigatons of carbon (GtC) per year from the atmosphere, while we now emit somewhere in the neighborhood of 35 GtC per year. Land use changes (afforestation, planting more carbon sequestration efficient crops, &c.) are just not capable of operating at the scale that would be necessary to offset the fucking insane amount of CO2 humanity is pumping into the atmosphere. Realistically, we could maybe achieve something on the order of 0.8 GtC removal by 2030 with land use changes like this--it's not nothing, but it's not even 2% of what the projected CO2 emission increase is going to be over the same period. The amount of CO2 we're emitting as a global civilization (and the rate at which those emissions are increasing) is simply too high for these sorts of interventions to work on their own. Technological carbon capture and sequestration (along perhaps with enhanced weathering) are a bit more promising, but also unlikely to function at the scale (and cost effectiveness) that we'd need them to, at least in the short to medium term. The only reasonable path out of this mess we've created is to rapidly and aggressively decarbonize, which almost certainly also means dismantling global capitalism.

This report from the Royal Society is a good comprehensive overview of the costs and benefits associated with different geoengineering approaches, including carbon capture and sequestration via land use change.