Could a U.S. Website Be Copyrighted Under State Common Law?
» Email a colleague about this essay.
Websites generally can be under copyright. Some aren’t and some portions aren’t, but most of them can be. And most of them are protected by Federal copyright law. You don’t even have to register the work with the government or publish a copyright notice, although both steps are usually helpful. But, while I’m not a lawyer, I wonder if State copyright law wouldn’t apply to some content.
Specifically, the content that is constructed on the fly only after a user has requested it and is never stored or written to a medium by the hoster is never in a fixed medium (a CD or DVD is a fixed medium and probably magnetic and solid-state media that store past power-off are, too). That’s a key to Federal copyright law. Federal copyright law pre-empted State copyright law, but it pre-empted it only for works in fixed media. For all other works, State law still rules. State law on point is generally common law, the law we imported from England in colonial times and partly amended here since then (England no longer decides it for us). There are also some State statutes.
A speech created in the head and given from a tree stump or a soap box, with or without electrical amplification but without a written script, is likely under State law. A transcript prepared afterwards might be a derivative work, thus maybe irrelevant to this issue. The derivative work in a fixed medium could get Federal protection. But a bard singing original lyrics might look to State law.
Computers contain firmware and software and both are generally in fixed media, such as CPU processor chips and hard drives, bringing firmware and software generally under Federal coverage.
But computers generate output. The output might be unrecorded by the hoster. Artificial intelligence may generate massive amounts of complex output that is highly useful, highly valuable, and copied with relative ease, but is not recorded by a hoster.
A program might have lists of possible outputs. Eliza, which was an early effort at artificial intelligence, a model of the Turing machine by which a user could be fooled into thinking the computer (if hidden) is really a human, had lists of words it could use in certain contexts. Those lists would be stored with the program on a hard drive, thus in a fixed medium, so Eliza would be subject to Federal copyright. But the actual outputs might not be fixed anywhere. If a program (whether or not Eliza) could construct a three-word sentence from a choice of ten subjects, ten verbs, and ten objects, that alone could generate a thousand possible sentences. Eliza could conduct a dialogue of some length. Probably millions of dialogues, just counting one side, were easily possible. If the hoster transmits the output without storing it, then no fixed media apply, and State law could control.
Retailers with websites allowing customers to compare a few products may come under this as well. If a small retailer has just a hundred products, and a customer can compare two to five products and even without varying the order for viewing, the number of comparisons may need a calculator. It’s easily in the billions. That’s too many to store in advance waiting for requests. Instead, the computer waits until someone asks and then only the desired comparisons are created on the spur and those are promptly transmitted to users without storage.
A trivial case is of the content inside a microprocessor chip. Consider this problem: 10000000000.1 + 10000000000.1 + 10000000000.1 = 30000000000.3 Easy for you to do. Complicated for a computer to do. We know how it’s done inside most computers. The instructions for a microprocessor are available. Maybe they’re a trade secret, but someone with access could use them to write out all of the steps. Most of those steps would be observable only in flowing electricity and never stored, only the output to a register and the output to the BIOS, the operating system, and a user’s application being stored. So some of those steps would be under State copyright. But this has little practical value, since most microprocessors probably are not able to support observing all of the steps. That would make copying a specific step impossible. The step could be predictable with absolute accuracy but still not observable. In that case, I think infringement is impossible. If so, copyright lacks practical value.
Recording by a recipient is generally irrelevant. If the recipient is someone else, they don’t have the copyright, they can’t claim the copyright as their own, and they can’t act on the hoster’s behalf to claim the copyright for anyone. What matters for the creator’s copyright is what the creator does about it. So, whether or not the recipient records the previously-unrecorded content onto a fixed medium when the creator didn’t, State copyright may apply.
Registration with the States is, as far as I know, not required.
It’s rare these days to find any new work that comes under State law. But I guess it’s possible that it could apply to the Internet under rare circumstances, and, with the growth of AI, not-so-rare circumstances.
