@1cig; Are you seriously arguing that AI and other forms of automation are never going to be a threat to anything more than low-skill labor? IBM's Watson has been around for years, and is only improving over time. Other companies are developing similar software. While certainly not the most real-world scenario, you might remember Watson defeating two of Jeopardy's most successful contestants with a score higher than both combined, and that was six years ago.
As for the legal profession, it's not all trial and drama. In fact, trial lawyers are typically in the minority in their firms. A vast amount are employed doing the trench work of discovery, research, etc. Discovery, particularly ediscovery, is entirely within the competence of any Watson-like AI. Which do you think is going to be more efficient, a room full of humans scrolling through screen after screen of text, each only able to maybe skim a few pages a minute, or a software that can multi-task between any number of documents simultaneously, and which does not need to take a coffee or lunch break? Then take legal research. Westlaw and LexisNexis have been fully digitized for well over a decade, making them ideal platforms by which to be parsed by AI.
Sure, there will still be a need for some attorneys, like the aforementioned trial lawyers, but the circumstances are not all that dissimilar to the retail implosion, perhaps just deferred a decade or two. Retail will always need some overpaid executives to make various decisions that typically make everything worse for everyone, and perhaps a tiny amount of people in the end to do some niche functions that are unsuited for automation.
The point, as already mentioned by another poster, is that no field is truly safe from automation. Even where humans are still needed, there will always be emerging technology that minimizes that need. Some might say that the tertiary industry was the last fallback for large-scale employment. The primary and secondary had already experienced massive automation by relatively dumb robots and machines decades ago. If society does not choose wisely about what to do about this problem, we may see the wealth gap explode to levels not seen since feudalism.