The NLRB is gearing up to implement a new rule which would establish that graduate students, even when acting as teaching assistants, are not "employees" of the university and therefore are not entitled to labor law protections (such as the right to unionize). This is the latest swing in an ongoing battle over that issue at the NLRB, which has largely tracked partisan politics: grad students first won the right to unionize under a Clinton-appointee controlled board in 2000, lost it under the Bush administration in 2004, gained it back under Obama in 2016, and are poised to lose it again under Trump in 2019.
The logic of the argument that grad students are not employees is that we are primarily students, and our teaching roles are extensions of our role as students. I'm somewhat uniquely positioned to address this, as I became a graduate student after serving as a (non-permanent) faculty member at two universities, and thus was able to compare whether and to what degree serving as a graduate student instructor (Berkeley's term for a teaching assistant) differed from teaching as a faculty member.
Here's my answer: there's virtually no difference.
There are many things I've done at UC-Berkeley as a "student". Most obviously, I enrolled in classes for credit. I wrote a master's thesis, and will write a dissertation. In these roles, my relationship to the departmental faculty is pedagogical -- they are there to teach and mentor me, and I interact with them in effectively the same way as I did to my teachers as an undergraduate or a secondary school student.
But as far as Berkeley is concerned, I do not teach classes to be taught how to teach. I teach classes because Berkeley needs people to teach classes, and they've decided I'm qualified to do it. As a GSI, both my day-to-day and semester-long routines are effectively the same as when I was a faculty member. I build my own lesson plans, hold office hours, give lectures, grade papers, write letters of recommendation, and often times create assignments. Nothing in how GSI work is structured remotely resembles any pedagogical practice akin to the professor/student relation that exists when I, say, submit a dissertation draft. As a GSI, I receive effectively no mentoring or even oversight by faculty members (indeed, in three of the four classes I've taught as GSI, the lead instructor was another graduate student). The only substantial difference between being a lecturer at the law school and a GSI in the political science department is that in the latter case I do not have control of my class's overall syllabus -- but all this means is that I'm a worker underneath a boss. To the extent that Berkeley has taken an interest in teaching me how to teach, it's done so through a class in pedagogy. It is an accident of structure that my GSI responsibilities come attached to my studies as a PhD student -- they are scarcely different in form than what I would be doing if I was an adjunct instructor.
This ruling will not directly affect me (it only applies to private universities). And I say all of this as someone who has had a sometimes adversarial relationship with my own graduate student union. But that doesn't change the obvious fact that when I teach, I'm engaging in work, not study, and the law should therefore treat me as a worker, not a student. The determination that graduate students serving as teaching assistants are not acting as "employees" but as "students" is patently absurd to anyone who has ever worked as TA.