Well, in principle they can via two photon absorption, but this process is usually not very likely as it is a nonlinear effect. Those electronic excitation processes happen very fast, somewhere between the attosecond and femtosecond range, so two photons need to "hit" the electron in a very short time window somewhere between 10^{-15} and 10^{-18} seconds.
In an hypothetical example you now have a monochromatic light source of 10 mW power (this is already more than one of the HeNe lasers in my lab has), which emits photons of 1 eV energy (somewhere in the infrared), but you need exactly 2 eV to liberate the electrons. A quick calculation shows, that there are about 6,2 * 10^{16} photons emitted per second, so the number of photons arriving within a femtosecond is about 60 photons. Now you also have to consider, that the time window is smaller than a femtosecond, the beam will be very large compared to the cross section of the electron, conservation rules apply (spin, for example) and by far not every photon will interact with the electron. So even with a VERY optimistic estimate, you will have a few thousand two photon absorptions per second and a few liberated electrons.
So the reason, why the high energy photons work better, is that you have to compare the number of photons arriving within a certain time window (high energy case) to the number of photon pairs arriving within a certain time window (low energy case). So unless you go to very, very high intensities, liberating electrons with low energy photons won't work.