To Bill's point, as architects, we know golfers assume risk and for their own play, responsibility. [size=78%]I[/size] have taken golf specific seminars on golf ball strikes and liabilities as it relates to course design.
Generally, our designs need to meet the maddeningly vague standard of "preponderance" of golf balls off adjacent properties and fairways. We tend to set higher standards for adjacent roads (where a missed shot could disable a driver causing multiple deaths) then adjacent land uses (where an adjacent ball strike would likely kill or injure just one person) and lastly the golf course. Even there, I am cognizant of adjacent fairways, where golfers are spread out, which get less consideration than adjacent greens or tees where 4-8 people (or more on certain tees) will be congregated, perhaps increasing the potential for injury.
Design tools include pure distance separation, maximum angles of dispersion (i.e., the "safety cone"), and other mitigation, like trees, fences and netting, up slopes at the perimeter to kill roll, vertical differences, etc. Also to be considered is the actual visual line of play, which in the case of open doglegs, isn't always down the intended line of play as marked on plans. I saw (and declined the opportunity to be involved) in a legal case where the homeowner was getting pelted even though he was close to a suitable distance from the fw centerline. The only problem was, a huge tree blocked play to the center of the fw and he was very close to the angle golfers had to play. Just an example.
Our charge is basically to avoid repetitive injury situations. Or, as one lawyer told me, "The first injury is on the house." At some point, if there are repeated incidents (including near misses) the course is probably legally smart or even obligated to provide some kind of mitigation. This is just my observation, but I suspect over time, the law favors those who are struck more and more. Even houses built decades after the golf course now presume their safety is the golf course's responsibility, even at they build 10,000 SF of a house in a 6,000 SF bag, and closer to the course than it "needs to be."
Given all the old courses that are built to what we now presume to be obsolete standards for new courses, the cases get interesting. Testifying that "there are thousands of courses out there that are tighter" isn't really a strong defense. So, as Bill notes, given the variety of laws across 50 states, and the unique nature of each situation, there are likely to be some, shall we say, interesting results if these things go to court.
As you can probably guess from my more scientific bent towards some aspects of design, I have taken a lot of shot data to set my personal design standards, but would never publish those, as the entire industry is reluctant to set any kind of safety standard for
design, given how random most ball strikes are.