The Supreme Court of Canada ruled that hosting a party does not make you legally liable if one of your guests drives drunk and hits somebody.
The result is a bit tragic. The dilemma the tort system has faced for a century is that it is the only close-to-adequate compensation scheme for the terribly injured (other than workplace-based first-party disability insurance), but it is supposed to be based on the fault of the defendant. But when that fault is most extreme -- drunk driving, for example -- the insurance companies claim breach of the conditions of coverage. This creates a desire to find some other insurer. The plaintiff lawyers here looked around, and found the home insurers of the party hosts.
The court was right, though. The benefit of policing our friends' liquor consumption is outweighed by the costs. A decision the other way would really be burying personal responsibility.
The way they got there wasn't too bad. The Anns framework was as useless as ever, and the Court didn't even do it right: it acted as if the issue was whether this injury was foreseeable to these hosts, but the question of duty of care surely depends on whether drunk driving accidents are, in general, foreseeable to social hosts, which they are (but surely that isn't enough). One of the good things about the Cooper revolution was that it made clear that this is not good enough. The other is that it correctly returns us to categories and analogies/distinctions. The Supreme Court made sensible distinctions between commercial liquor establishments and party hosts, including pointing to regulation and the incentive bars have to over liquor their patrons.
The people who lose here are the victims of drunk driving accidents. And they lose because we insist on tying their compensation to the tort system, instead of developing a cheaper no-fault compensation system (or, at minimum, preventing auto insurers from walking away from drunk driving liability).