This is the article:
http://www.nytimes.com/2009/06/24/business/energy-environment/24geotherm.html
And this is one very nice and informed comment from a PhD student that puts things into the right perspective:
1) The major difference between enhanced geothermal and standard oil and gas extraction is the high pressures used to create geothermal reservoirs. Oil and gas are normally extracted from porous sedimentary rocks (10-30% void space), but the heat fluxes required for geothermal are found in volcanic rocks with less than 1% void space. The rock must be fractured to allow water to flow through it, and it is this process that causes earthquakes. The earthquakes can continue for months after the pressure has been released because the Earth is a dynamic place and not all of the stress transfer is instantaneous; in fact, the later-occurring quakes are usually governed by stress changes caused by diffusion (slow filtering through the limited porosity due to pressure gradients) rather than advection (the injection velocity/pressure, which causes the immediate earthquakes.
2) There is no danger of creating “new tectonic plates”; the tectonic plates are on a completely different scale than a single geothermal reservoir (1000s vs 1s of kilometers). The absolute worst case scenario would be changing the stress regime enough to reactivate an existing tectonic-scale fault like the San Andreas/San Jacinto etc., which is very unlikely due to the scale difference.
3) For risk assessment purposes, risk is the product of hazard (the event that might be triggered) and vulnerability. A project like this ends up going forward because when the hazard is large (i.e. it is very likely that you will trigger small earthquakes), the vulnerability is also low since it will only affect a relatively sparsely populated area immediately nearby. Obviously, if you happen to live in that area, that is not a reassuring argument. In the other extreme case, if the vulnerability is high (eg. a Landers earthquake-type rupture across most of SoCal), the hazard is low (the chance of re-activating the San Andreas is slim). In both cases, the product of the two works out to a small number, so the risk is deemed acceptable. The soft underbelly of risk assessment is that seismic hazard is fiendishly difficult to quantify given present methodology and observational capabilities.
4) Regional seismic hazard and geothermal energy will continue to be coupled since the areas with the highest geothermal gradients (rate of temperature increase with depth) tend to be tectonically active (young or recently deformed) crust. However, the potential benefits to figuring out geothermal energy are myriad - clean, reliable, domestic supplies - and I think it’s worth pursuing. An exciting tangential scientific opportunity also exists: if we can figure out the processes that control microseismicity in reservoirs, we will be much closer to predicting and someday mitigating the types of earthquakes that make headlines on a more regular basis.
5) This may come as a surprise, but the largest geothermal producer in the world is the oil giant Chevron; they have several enormous traditional geothermal projects in Indonesia.