Calling it "freedom" could not be farther from the truth. You don't get to work on whatever you want unless you want to pay your own way. Stray a little bit out of your advisor's comfort zone and prepare to be put in your place. There will be plenty of interruptions, and you will spend countless hours digging holes and filling them back up as you carry out orders from emails that you get at 2am from your advisor.
That salary is abysmal when you add up 80-hour weeks over 5+ years. Add in the risk of failing out of the program or quitting, and you would have to (in many cases) be a fool to take the deal.
See here for a realistic picture of what it can be like to go through with a PhD in computer science: https://news.ycombinator.com/item?id=8720640
I hear about people with dysfunctional phd experiences and one of my friends is experiencing one right now. To be honest, the students are to blame at least partly for this. IMO the people who get most exploited are those who are most desperate for a phd. Like someone said in the reply to the linked comment, you shouldn't tolerate abuse, you need to do your research before you sign on the dotted line and you should be prepared to walk away if things do go south.
Sure, there are jerks in academia, but I've worked for a few different tech companies and there are plenty of jerks there too. We never think twice about getting out of a shitty job situation, but somehow people don't do the same in a phd program.
I also don't know what people are talking when they mention 80 hour workweeks. I do something like 40-60 a week. Some weeks are closer to 30 than 40, and the 60 hour weeks are rare. The people who do 80 hour weeks don't seem to know when to stop working, and they don't seem to be able to accept that there are others who are way smarter than them and working harder isn't going to make up for that. I realized very quickly in my phd that there were people here who were in a different league from me in terms of ability. If I'd tried to compete with them, I would have lost both the rat race and my peace of mind, so I didn't. What I do know is that my work is interesting and important and clearly people appreciate it, I'm satisfied with that. I'm sure I could be winning more awards and publishing 4 papers a year instead of 2 if I worked 20 hours more every week, but really, I don't want to do that.
I'm not arguing that there's total freedom to go to grad school and do what you want. I'm arguing that no one is forcing anyone to be in grad school doing something they don't want.
Ultimately, it is geared around your two major milestones of confirmation (at 1 year) and submission (at 3-4 years). As your supervisor is not an examiner for your thesis at submission, this means that you are very much in control of what you do and write in your PhD. Your supervision team are your board of advisors, not actually your boss. (Though usually their advice is good, well-intentioned, and worth heeding.)
This does mean that in grants, we have to be a little careful what we say will be done by PhD students, because it's always possible they'll head in a different direction instead!
Actually, as I understand it (I'm a fairly junior academic) most Australian academics prefer capable independently-minded students. That's not just because we're nice people, but because that's how the incentives work -- we're all time pressured, so students who we can be confident will produce a good thesis with just a bit of direction and advice are like gold dust. Ones that we need to give more direction to are also good -- we've been through the process a few times and can design some PhD-shaped research for you to do. Not-so-capable independently-minded students I guess could be a bit trickier, as it could take quite a lot of time and effort to get their research into a shape that will be approved by the external examiner. (I'm yet to encounter the last category though.)
But I'm yet to hear of a 2am email that a student must answer at 2am.
But then, Australian PhDs tend to be 3-4 years' pure research, rather than 5+ years including lots of TA work, so there are a few differences in the system.
Plenty of excellent engineers take unpaid time off work between jobs for the opportunity to work on personal projects without interruption.
Point is, 30-40k/yr ain't bad pay for that sort of freedom.
EDIT: I've been in a PhD program. The reply to this comment DOES NOT describe a normal or healthy advisor/advisee relationship.
Ultimately, there should be a point in every PhD where the student carves out a project and works on that project, and it should probably happen before the student is spending most of their time on research. The student shouldn't have to play politics to convince people to allow him/her to work on their own ideas, as long as those ideas have scientific value.
If you find yourself consistently working 80 hour weeks and answering 2am emails, switch advisors or switch programs.
If you're just working on your advisor's stuff without any autonomy and/or don't have any say in the direction of your or your lab's research, switch advisors or switch programs.
A lot of the "bad press" PhD programs get aren't indicative of a healthy PhD and are not as common as they're made out to be. And never necessary. It's usually a combination of an abusive professor and a niave student. Which, BTW, happens in industry as well ("death marches" etc)
A phd isn't a 40 hour work week kinda gig, but it also shouldn't be abusive. But in general, it's really simple: don't let people abuse you, and if you're taking a pay cut in return for freedom, be damn sure you demand the freedom.