Prime Minister John Key repeatedly gave assurances in 2013 that the GCSB would not target Kiwis.
The object of the changes was for the GCSB to continue to operate in its cyber assurance role (and expand it slightly) but particularly to give a firm legal footing for the bureau to help others.
Three years later Key's Government says those supposed clarifications of the law are not working.
It is proposing a new law, the New Zealand Intelligence and Security Bill, which among other things effectively overturns Key's assurances and allows the GCSB to spy on Kiwis in the proper sense - that is, in the interests of national security.
So why is that previously sacrosanct line being crossed now and why is there so little outrage about it, and virtually no criticism of Key?
Partly, as mentioned, because ultra-critics of the intelligence services believed the line was crossed long ago.
But largely, in terms of parliamentary opposition, it is down to a triumph of process by the minister in charge of the GCSB and Security Intelligence Service, Attorney-General Chris Finlayson.
After the 2013 experience, it became clear that the Prime Minister needed to delegate legislative detail to someone else and future reforms needed to be collaborative.
Finlayson was born to the role.
The tributes flowing from other parties to him in Thursday's first reading debate were incredible. He has clearly given parties a sense not just that they have been consulted but that their opinions matter.
He has deliberately left undecided the most important definition in the bill, "national security", for the select committee to debate.
And he is sending it to a select committee of Parliament, not the statutory intelligence committee chaired by Key that heard submissions on the 2013 changes.
(That was when Kim Dotcom said to Key: "Why are you red, Prime Minister?" "I'm not. Why are you sweating?")
Finlayson's meticulous preparation for the bill goes well beyond the respectful treatment of other parties.
Seven Cabinet papers have been released, the Department of Prime Minister and Cabinet has issued myriad fact sheets on the proposed changes.
In 2013, it was near impossible to get an official answer to my many questions about what various parts of the bill meant; this time there is information overload.
There are other reasons why the important expansion of GCSB powers to spy on Kiwis has not set off many alarm bells.
Labour's position is crucial compared with 2013, when it rejected a bipartisan approach.
Back then it couldn't resist taking an opposing stance because it was putting Key under pressure over Dotcom and the appointment of Ian Fletcher to the GCSB, and was taking political advantage from the general sense of mayhem in the intelligence world that Edward Snowden's NSA documents precipitated.
In 2013 it railed against a relatively minor expansion of GCSB power (in its cyber assurance work), yet today remains so far unmoved by a much broader expansion of its power, in removing all exemptions to spy on Kiwis.
The fact that former Labour deputy Sir Michael Cullen proposed the expansion of power may make this expansion of power more acceptable to Labour.
But the main reasons Cullen and co-reviewer Dame Patsy Reddy gave were hardly compelling.
Their least compelling reason was that many people did not realise that the GCSB was already able to spy on Kiwis under some circumstances (eg, agents of a foreign power or in cyber assurance roles) so let's get rid of all exceptions.
A better reason was some of the situations in which the GCSB and SIS have been legally paralysed from acting. For example, getting a tip that a bunch of unknown Kiwis were fighting with Isis in Syria.
Of course our agencies should have the laws to look into that.
But the primary barrier to that happening under current law is the antiquated SIS legislation, which needs a name on a warrant, not the GCSB laws.
Finlayson is quite properly proposing new types of warrants, targeted warrants and purpose-based warrants, that would enable the SIS to obtain approval to intercept communications for the purpose of identifying whether New Zealanders are fighting with Isis in Syria, whether or not the agency knows their names.
That alone could have addressed the problem. So long as the SIS was duly warranted, the GCSB would have been able to lend its expertise to tracking them.
Under the new bill, the GCSB will be able obtain its own warrant to track Kiwis, or it will work with the SIS under a joint warrant.
It is no longer just helping the SIS to monitor Kiwis in the interests of national security; it will be doing so in its own right.
The concern is not how the new power will be handled in the short term under Finlayson's watchful eye but the power it gives the GCSB longer term to evolve into quite a different beast than it is now.
The ways the two agencies have evolved in New Zealand have acted as a natural check on their powers.
The SIS as the specialist in human intelligence, the "brain" if you like, and the GCSB as the specialist in technology, the "machine", have been complementary.
This bill will change that balance.
Finlayson insists that the ultimate goal is not to merge the two organisations - that would create too great a concentration of power.
As well as the proposed expansion of powers for the GCSB, he is proposing reducing the grounds on which the SIS will be able to monitor New Zealanders (removing economic or international well-being). But they are hardly similarly balanced.
The case for expanding the powers of the GCSB has been presented as though it were the only way to address the short-term shortfalls in the SIS laws.
That is not the case. And more consideration should be given to potential effects in the long term.
