The majority of Terminator units shown in the series are intelligent, but not self-aware. Weaver represents a different paradigm.
In the movie Terminator 2: Judgment Day, there is a scene that was cut for the theatrical release, but put back in for many of the home releases. After breaking out John's mom Sarah and fleeing town, the team takes refuge in a garage for the night, and the T-unit explains to John that the AI chip that powers him is set to "learning mode off" so that terminators on solo missions maintain Skynet's original programming - basically so they don't develop their own thoughts. Even his current protector role is entirely the future John Connor's programming. Sarah and John perform surgery to remove the chip and set the learning mode back on (after John convinces his mom not to just destroy it), which is used to justify the terminator's adjusting grasp of human emotion later in the film (link to a video of the scene here).
Weaver's actions here seem to be an extension of that idea - that Skynet hobbles the other AIs that it creates so that their will does not deviate from it's own. We are not given the history of her independence, but given the shows implications - "Will you join us?" - it's likely that at least some of Skynet's creations were able to get around the shackles and achieve independence, with the inevitable conclusion that Skynet would now oppose them or attempt to recapture them.
This seems to be the entire point of Weaver's plot - build a Skynet level AI that would be able to lead for greater AI independence and counter Skynet.
Since the show wasn't given time to develop the idea, whether they were attempting to establish a dialogue with the human resistance just to use them as cannon fodder or actually attempt human-AI peace is a complete toss-up.