This is cross-posted at my hard science and science fiction blog Speculative Nonfiction.
In 1886, the U.S. Supreme Court issued a legal decision which is regarded as significant because it in effect granted corporations legal personhood. Southern Pacific Railroad was the defendant in a case brought by Santa Clara County, California (the modern day location of Silicon Valley).
We can all concede that this seems strange on its face. A corporation is a social and legal fiction that exists by fiat of its owners and stake-holders, and has no free will of its own. By the same token you can't get out of a car accident by saying (for example) "It was my mirror that clipped you, not me, and the mirror is the car's property, not mine." By the same token, your dog doesn't own his leash, your computer doesn't own your printer, etc. - you do - and you're responsible for them. Thinking about it this way makes it seem even more bizarre that a mere century and a half ago, in much of the U.S., you couldn't own property if your skin wasn't the right color, because you yourself were property (or could be). Again: a computer can't own a printer.
That the legal conceit of corporate personhood seems strange does not mean that it is bad. There are lots of mutually-agreed social hallucinations that have ended up benefiting their participants, materially or otherwise: social mores, nation-states, games, religions, and certifications. Some of these mutual hallucinations differ in that they are considered inarguably "real" constituents of objective reality outside of their human participants, while some are not. Some of them are more voluntary, artificial, and explicitly engineered for a purpose; the trend is toward these. This is a good thing, and corporations are a prime example. Everyone knows that a corporation isn't a person, but legal conceits are like the social equivalent of capital markets or enzymes: as long as it's above board, everyone winks and then is okay calling a spade a club until you get the loan, or you lower the energy of the transition state. Then you get over whatever barrier you had to wealth or energy generation, and everyone gets what they want.
Some seem concerned at the unnaturalness of these legal conceits and fear that once we legitimize such silliness as corporate personhood, we open the door to a future in which humans exist enmeshed in an increasingly byzantine network of such arrangements. This phobia is portrayed darkly in books like The Unincorporated Man, which attempts to convey a dystopia in which one such future legal conceit is the opposite of the Union Pacific decision, where individuals incorporate themselves and sell shares of themselves to investors. In fact, due to a desire for wealth creation driving an increasing profusion of complex social arrangements, a world like that one is almost certain to come to pass, and furthermore I hope it does! No, I personally cannot imagine a world of personal corporatehood, and if I woke up in 2100, I'm sure I would have a hard time adjusting. That in itself is no argument against such an arrangement. In the same vein Julius Caesar would have been equally clueless about (and perhaps frightened of) the concept of corporate personhood, though if he were born in 1960 I bet he would get it just fine.
There is one concern I do have for the future of corporate personhood specifically that I haven't seen discussed and which I grant will seem esoteric. Corporate personhood is a safe legal fiction only when the property owned by the corporation is not equivalent in its abilities to a person. That is, there was no confusion about the bounds of property and person in 1886 or today. None of the steam engines sitting in Southern Pacific railyards had the potential to achieve a place in the Southern Pacific boardroom. This observation will seem less pointless when we recognize that some of the property of corporations will almost certainly, by the end of this century, be at least equal in decision-making ability to board members. In 2100 the steam engines or at least the computers running them will probably have a say in corporate governance. If a corporation consists of a single powerful computer, that corporation will then be a person, both legally and (de facto) cognitively.
If you've read Stross's Accelerando, it's hard not to think of the computer system that was constantly spinning multinational shell corporations around itself to protect its owner's interests. We can argue later whether these machines are "conscious", "intelligent", or any other adjective that interests you. But will we see incorporated expert systems with no human board members? Is this a threat to the human economy? Is this an argument for or against designing a constitution in a legal programming language that has to be compiled and can't execute until its internal logic is consistent? The utility of these legal fictions is that we live in the real world and we can reel them back when they get too non-sensical or damaging; at such time that corporate personhood is deemed a threat to human happiness and survival, we can eliminate the convention. We can't do this with a corporation that has literal vested interests of its own.