The fundamental idea is that if enough people sign, management at tech companies will have to make decisions to comply with unethical government orders knowing that the discovery of their compliance (which is inevitable) might cause significant retention problems.
In practice, developers work on those databases in Google and Facebook. They'll be "ok" because they don't know they are abused. Other engineers in those companies will work on giving the government access to it (as Snowden documented), and they too will not see any abuse so they are ok. The government gets that information and does what it wants with it. It might be abusing it right now. (There have been many deportations under Obama, for example, with little uproar; it's possible tech databases helped in those.)
The pledge is toothless. To be actually effective, it should have called on people to not work at Google and Facebook.
The real problem is the existence of the databases, and government access to them. Both of those things are facts right now. The pledge ignores that.
What if the petition itself is abused as a source of information?