Pumping water into faults
That much at least was based on serious engineering proposals made in the 1960s. The idea being if you pump water into fault zones you can increase the pore pressure in the rocks to a point where they overcome the sticking pressure holding the fault closed. If you could control the pressure you might be able to allow the fault to move gradually rather than in one catastrophic jolt.
It was planned to deploy it around Los Angeles where the San Andreas fault group makes a near 90 degree turn and is locked in place by the northwards movement of the Pacific plate. There have been no large 'quakes in the area in most of historic times, so the fault is under enormous pressure.
The real problem is that we don't know nearly enough about how faults break. Generally when one part of a fault breaks it transfers some of its energy into adjacent sectors of the fault, if they were close to breaking you could trigger another earthquake. So the liability issues are huge.
Also, the number of 'quakes needed to destress a fault would be massive - you'd need tens of thousands of smaller shocks to produce as much energy as is probably already accumulated in the San Andreas near LA.
The theory came about because a link was noticed between the frequency of earthquakes in Colorado and the pumping of nerve gas wastes down a deep borehole. As more liquid went down, the frequency went up. We see the same correlation around geothermal power plants which return spent well water to the reservoir, in some oil and gas fields where fluids are injected to recover more produce, and around large reservoirs where water is being forced into faultlines.