The problem with this is that when compared with the brightness of the star the alien world is orbiting, city lights will be very hard to distinguish. Schneider et al. estimate that we'd need to build an array of telescopes with the collective area of 1.5 square kilometers. Even then we'd only be able to search for light pollution on exoplanets within 15 light years from Earth.
No, when searching for alien civilizations further afield we'd need to look out for some other indication of a civilization. What kind of pollution doesn't occur naturally, indicating a civilization is thriving?
The detection of chlorofluorocarbons (or CFCs) in an exoplanet's atmosphere would be a very good indicator of the presence of alien technology. CFCs do not occur naturally and they are strong absorbers of infrared light. Although a very powerful telescope and spectrometer would be required, the existence of CFCs could be detected by analysing the parent star's starlight through the exoplanet's atmosphere.
Although this is a very novel way of looking for ET, we'd be making a huge assumption that the alien civilization has made the same mistakes as mankind in producing CFCs in the first place. After all, CFCs make holes in our ozone layer, this makes CFCs bad, right?
Although CFCs are now banned on Earth (and they will eventually be removed from our atmosphere all together), say if alien civilizations don't need an ozone layer? (This is a tricky argument to present as ozone absorbs harmful ultraviolet radiation from the sun, thereby allowing life to evolve. But this is terrestrial life, perhaps "life as we don't know it" evolved differently to us.) Assuming CFCs are bad universally - simply because they eat our protective ozone layer - is another assumption that may not apply to an alien species.
Whether or not we ever detect the presence of an alien civilization, we are certainly giving the search our best shot. Personally, I wonder if we'll be lonely for much longer...
Source: New Scientist